intro_2023_exercise_8_solution.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyPZWXdDaUnoh5LQ9IBHqRtW",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/TurkuNLP/intro-to-nlp/blob/master/intro_2023_exercise_8_solution.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Sequence labeling with CRF\n",
        "\n",
        "This notebook provides an example of how to perform sequence labeling with a Conditional Random Field (CRF) model.\n",
        "\n",
        "(This notebook is based on this [notebook for CoNLL 2002 Spanish tagging](https://github.com/TeamHG-Memex/sklearn-crfsuite/blob/master/docs/CoNLL2002.ipynb))"
      ],
      "metadata": {
        "id": "RFGEAWhFZoYy"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "---\n",
        "\n",
        "## Setup\n",
        "\n",
        "We'll be using the [sklearn-crfsuite](https://sklearn-crfsuite.readthedocs.io/en/latest/) Python package, which wraps the [CRFsuite](http://www.chokkan.org/software/crfsuite/) CRF implementation.\n",
        "\n",
        "(Specifically, we'll install a fork that fixes a [versioning incompatibility issue](https://github.com/TeamHG-Memex/sklearn-crfsuite/issues/60) with the integration)"
      ],
      "metadata": {
        "id": "Klope-YOaLmm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install --quiet git+https://github.com/MeMartijn/updated-sklearn-crfsuite.git#egg=sklearn_crfsuite"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uhyqgRJHo5Rv",
        "outputId": "dd603221-b50d-40d7-f61a-a927c7c2c1ad"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "To take advantage of the integration of CRFsuite into the [scikit-learn](https://scikit-learn.org/stable/) machine learning library, we'll here be using parts of that library to support data loading and evaluation. (The basic functionality provided by this library is similar to what you could get from the `datasets` and `evaluate` libraries that you should know by now, and there's no need to try to memorize any of the specific details of these libraries.) "
      ],
      "metadata": {
        "id": "5tor3S1madzC"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "VbMIi8zaojau"
      },
      "outputs": [],
      "source": [
        "from itertools import chain\n",
        "\n",
        "import nltk\n",
        "import sklearn\n",
        "import scipy.stats\n",
        "import sklearn.metrics\n",
        "\n",
        "from nltk.corpus.reader import ConllCorpusReader\n",
        "\n",
        "from sklearn.metrics import make_scorer\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import RandomizedSearchCV\n",
        "\n",
        "import sklearn_crfsuite\n",
        "from sklearn_crfsuite import scorers\n",
        "from sklearn_crfsuite import metrics\n",
        "from sklearn_crfsuite.utils import flatten"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "---\n",
        "\n",
        "## Load data\n",
        "\n",
        "We'll use the CoNLL'03 English dataset, which you should be familiar with. We've placed the original data on dl.turkunlp.org, and can download with `wget` as usual: "
      ],
      "metadata": {
        "id": "0ctbbmzIbY6t"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!wget -nc --quiet http://dl.turkunlp.org/TKO_7095_2023/eng.train\n",
        "!wget -nc --quiet http://dl.turkunlp.org/TKO_7095_2023/eng.testa\n",
        "!wget -nc --quiet http://dl.turkunlp.org/TKO_7095_2023/eng.testb"
      ],
      "metadata": {
        "id": "xMHcKOGgb1Em"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "We'll use the NLTK `ConllCorpusReader` to read in the words, parts of speech and named entity \"chunks\""
      ],
      "metadata": {
        "id": "OxzH2EyoelQ4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_data = ConllCorpusReader('', 'eng.train', ['words', 'pos', 'ignore', 'chunk'])\n",
        "devel_data = ConllCorpusReader('', 'eng.testa', ['words', 'pos', 'ignore', 'chunk'])\n",
        "test_data = ConllCorpusReader('', 'eng.testa', ['words', 'pos', 'ignore', 'chunk'])\n",
        "\n",
        "# The comprehension here drops empty sentences used to denote document boundaries \n",
        "train_sentences = [s for s in train_data.iob_sents() if s]\n",
        "devel_sentences = [s for s in devel_data.iob_sents() if s]\n",
        "test_sentences = [s for s in devel_data.iob_sents() if s]"
      ],
      "metadata": {
        "id": "x-3Jt5QRdPuw"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "This reader returns a simple Python list representing the sentences, each of which is represented as a list of tuples of token features (form, POS tag, NE tag).\n",
        "\n",
        "Note that unlike in `datasets`, the data is here in \"human-readable\" string form rather than as integer IDs."
      ],
      "metadata": {
        "id": "KtI0SxUblWeK"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_sentences[0]"
      ],
      "metadata": {
        "id": "uVVNBV5yplN5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a9c7e153-5fc3-4332-da6b-dfc0b6a14a85"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[('EU', 'NNP', 'I-ORG'),\n",
              " ('rejects', 'VBZ', 'O'),\n",
              " ('German', 'JJ', 'I-MISC'),\n",
              " ('call', 'NN', 'O'),\n",
              " ('to', 'TO', 'O'),\n",
              " ('boycott', 'VB', 'O'),\n",
              " ('British', 'JJ', 'I-MISC'),\n",
              " ('lamb', 'NN', 'O'),\n",
              " ('.', '.', 'O')]"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "---\n",
        "\n",
        "## Create features\n",
        "\n",
        "Here a very basic function that returns explicitly defined features for a token"
      ],
      "metadata": {
        "id": "dGrCQrSkh9zS"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "```\n",
        "Initial           : micro avg       0.86      0.83      0.84      8603\n",
        "+ token.lower()   : micro avg       0.87      0.84      0.86      8603\n",
        "+ window_size=1   : micro avg       0.90      0.86      0.88      8603\n",
        "+ window_size=2   : micro avg       0.91      0.86      0.88      8603\n",
        "+ BOS and EOS     : micro avg       0.91      0.86      0.88      8603\n",
        "+ token prefixes  : micro avg       0.91      0.87      0.89      8603\n",
        "+ token suffixes  : micro avg       0.91      0.88      0.89      8603\n",
        "+ hyperparams     : micro avg      0.914     0.878     0.896      8603  \n",
        "```"
      ],
      "metadata": {
        "id": "BwNr__mV7bfm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def token_features(tokens, index):\n",
        "    features = {\n",
        "        'bias': 1.0,\n",
        "    }\n",
        "\n",
        "    window_size = 2\n",
        "    start = max(0, index-window_size)\n",
        "    end = min(len(tokens)-1, index+window_size)\n",
        "\n",
        "    if start > index-window_size:\n",
        "        features['BOS'] = True\n",
        "    if end < index+window_size:\n",
        "        features['EOS'] = True\n",
        "\n",
        "    for i in range(start, end+1):\n",
        "        o = i - index\n",
        "        token = tokens[i][0]\n",
        "        pos_tag = tokens[i][1]\n",
        "        features.update({\n",
        "            f'token[{o}].text()': token,\n",
        "            f'token[{o}].lower()': token.lower(),       \n",
        "            f'token[{o}].istitle()': token.istitle(),\n",
        "            f'token[{o}].isdigit()': token.isdigit(),\n",
        "            f'pos_tag[{o}]': pos_tag,\n",
        "        })\n",
        "\n",
        "    for i in range(2, 4):\n",
        "        features.update({\n",
        "            f'token[0].text()[:{i}]': tokens[index][0][:i],\n",
        "            f'token[0].text()[{-i}:]': tokens[index][0][-i:],\n",
        "        })\n",
        "\n",
        "    return features"
      ],
      "metadata": {
        "id": "orTF5hBlpo0z"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Some convenience functions to get features for all tokens in a sentence as well as all labels and all token texts."
      ],
      "metadata": {
        "id": "vggsScuniQ8m"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def sentence_features(tokens):\n",
        "    return [token_features(tokens, i) for i in range(len(tokens))]\n",
        "\n",
        "def sentence_labels(tokens):\n",
        "    return [label for token, postag, label in tokens]\n",
        "\n",
        "def sentence_tokens(tokens):\n",
        "    return [token for token, postag, label in tokens]"
      ],
      "metadata": {
        "id": "7asTewAKiNDW"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Check features for a single token"
      ],
      "metadata": {
        "id": "cBjEVBS7l6dR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "sentence_features(train_sentences[0])[0]"
      ],
      "metadata": {
        "id": "dk5c3rKcqBu0",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "13517c86-487e-4994-e91a-79a8d718ddfe"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'bias': 1.0,\n",
              " 'BOS': True,\n",
              " 'token[0].text()': 'EU',\n",
              " 'token[0].lower()': 'eu',\n",
              " 'token[0].istitle()': False,\n",
              " 'token[0].isdigit()': False,\n",
              " 'pos_tag[0]': 'NNP',\n",
              " 'token[1].text()': 'rejects',\n",
              " 'token[1].lower()': 'rejects',\n",
              " 'token[1].istitle()': False,\n",
              " 'token[1].isdigit()': False,\n",
              " 'pos_tag[1]': 'VBZ',\n",
              " 'token[2].text()': 'German',\n",
              " 'token[2].lower()': 'german',\n",
              " 'token[2].istitle()': True,\n",
              " 'token[2].isdigit()': False,\n",
              " 'pos_tag[2]': 'JJ',\n",
              " 'token[0].text()[:2]': 'EU',\n",
              " 'token[0].text()[-2:]': 'EU',\n",
              " 'token[0].text()[:3]': 'EU',\n",
              " 'token[0].text()[-3:]': 'EU'}"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Get features and labels for training and development data"
      ],
      "metadata": {
        "id": "qMa3tJbHjThA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train = [sentence_features(s) for s in train_sentences]\n",
        "y_train = [sentence_labels(s) for s in train_sentences]\n",
        "\n",
        "X_devel = [sentence_features(s) for s in devel_sentences]\n",
        "y_devel = [sentence_labels(s) for s in devel_sentences]"
      ],
      "metadata": {
        "id": "THQte3hTqDbR"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "---\n",
        "\n",
        "## Train and evaluate model\n",
        "\n",
        "Instantiate CRF. For details on the model hyperparameters, see the [CRFsuite documentation](https://www.chokkan.org/software/crfsuite/)."
      ],
      "metadata": {
        "id": "Su2nWOQ6jZUs"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "crf = sklearn_crfsuite.CRF(\n",
        "    algorithm='lbfgs', \n",
        "    c1=0.1, \n",
        "    c2=0.1, \n",
        "    max_iterations=100, \n",
        "    all_possible_transitions=True\n",
        ")"
      ],
      "metadata": {
        "id": "v-Jhj_qkqUzM"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Train the model"
      ],
      "metadata": {
        "id": "f2vguONKjnLJ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "crf.fit(X_train, y_train)"
      ],
      "metadata": {
        "id": "ZSUtNbIGjQI-",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 92
        },
        "outputId": "2016bb2e-9c80-4ac2-8053-cd21ac194f82"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "CRF(algorithm='lbfgs', all_possible_transitions=True, c1=0.1, c2=0.1,\n",
              "    max_iterations=100)"
            ],
            "text/html": [
              "<style>#sk-container-id-1 {color: black;background-color: white;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>CRF(algorithm=&#x27;lbfgs&#x27;, all_possible_transitions=True, c1=0.1, c2=0.1,\n",
              "    max_iterations=100)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" checked><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">CRF</label><div class=\"sk-toggleable__content\"><pre>CRF(algorithm=&#x27;lbfgs&#x27;, all_possible_transitions=True, c1=0.1, c2=0.1,\n",
              "    max_iterations=100)</pre></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Get set of POS tags for evaluation, excluding the \"Out\" tag `O`"
      ],
      "metadata": {
        "id": "SFIp-jnEjuUi"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "labels = list(crf.classes_)\n",
        "labels.remove('O')\n",
        "labels"
      ],
      "metadata": {
        "id": "xrPY6i55qXaJ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "16e9c990-2532-490a-be96-2ec4c809dc0c"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['I-ORG', 'I-MISC', 'I-PER', 'I-LOC', 'B-LOC', 'B-MISC', 'B-ORG']"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Predict labels for development set"
      ],
      "metadata": {
        "id": "OBzTAFh4j4bF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "y_pred = crf.predict(X_devel)"
      ],
      "metadata": {
        "id": "-m0c_Lw8qbNJ"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Evaluate predictions against gold standard development set labels"
      ],
      "metadata": {
        "id": "cjRLXER_j9mu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "sorted_labels = sorted(labels, key=lambda name: (name[1:], name[0]))\n",
        "\n",
        "print(sklearn.metrics.classification_report(\n",
        "    flatten(y_devel),\n",
        "    flatten(y_pred),\n",
        "    labels=sorted_labels\n",
        "))"
      ],
      "metadata": {
        "id": "NSEiOFvlqeKL",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8e6cc238-ccaf-4dd6-8e93-3ffecc58a90e"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "       B-LOC       0.00      0.00      0.00         0\n",
            "       I-LOC       0.92      0.88      0.90      2094\n",
            "      B-MISC       0.00      0.00      0.00         4\n",
            "      I-MISC       0.91      0.83      0.87      1264\n",
            "       B-ORG       0.00      0.00      0.00         0\n",
            "       I-ORG       0.87      0.83      0.85      2092\n",
            "       I-PER       0.93      0.93      0.93      3149\n",
            "\n",
            "   micro avg       0.91      0.88      0.89      8603\n",
            "   macro avg       0.52      0.49      0.51      8603\n",
            "weighted avg       0.91      0.88      0.89      8603\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "---\n",
        "\n",
        "## Analyze classifier\n",
        "\n",
        "We can look at `transition_features_` to analyze the transition weights learned by the model:"
      ],
      "metadata": {
        "id": "U7dSD_gUobjY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from collections import Counter\n",
        "\n",
        "def print_transitions(trans_features):\n",
        "    for (label_from, label_to), weight in trans_features:\n",
        "        print(\"%-6s -> %-7s %0.6f\" % (label_from, label_to, weight))\n",
        "\n",
        "print(\"Most likely transitions:\")\n",
        "print_transitions(Counter(crf.transition_features_).most_common(10))\n",
        "\n",
        "print(\"\\nLeast likely transitions:\")\n",
        "print_transitions(Counter(crf.transition_features_).most_common()[-10:])"
      ],
      "metadata": {
        "id": "bjtNiBFrqhyZ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "59826783-b9b4-4a17-dcdf-593d3b7674cc"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Most likely transitions:\n",
            "I-MISC -> B-MISC  3.727991\n",
            "I-ORG  -> I-ORG   3.661494\n",
            "I-LOC  -> I-LOC   3.314899\n",
            "I-MISC -> I-MISC  3.063793\n",
            "I-PER  -> I-PER   3.044030\n",
            "I-LOC  -> B-LOC   2.936341\n",
            "B-ORG  -> B-ORG   2.813874\n",
            "B-MISC -> I-MISC  2.259211\n",
            "O      -> O       2.077222\n",
            "O      -> I-PER   0.734798\n",
            "\n",
            "Least likely transitions:\n",
            "I-MISC -> I-ORG   -1.980208\n",
            "I-ORG  -> I-MISC  -2.190352\n",
            "I-MISC -> I-LOC   -2.605446\n",
            "I-LOC  -> I-ORG   -2.804579\n",
            "I-PER  -> I-MISC  -2.833085\n",
            "I-PER  -> I-LOC   -3.140030\n",
            "I-LOC  -> I-PER   -3.397410\n",
            "I-ORG  -> I-PER   -3.611332\n",
            "I-PER  -> I-ORG   -3.990547\n",
            "I-ORG  -> I-LOC   -4.026007\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Similarly, we can look at `state_features_` to look at the probabilities of the explicit features we introduced:"
      ],
      "metadata": {
        "id": "bKYDOb4pol0x"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def print_state_features(state_features):\n",
        "    for (attr, label), weight in state_features:\n",
        "        print(\"%0.6f %-8s %s\" % (weight, label, attr))    \n",
        "\n",
        "print(\"Top positive:\")\n",
        "print_state_features(Counter(crf.state_features_).most_common(10))\n",
        "\n",
        "print(\"\\nTop negative:\")\n",
        "print_state_features(Counter(crf.state_features_).most_common()[-10:])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CeOjznEsm5-G",
        "outputId": "cde85263-da18-4ce5-ed1a-5417f69b0d4b"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Top positive:\n",
            "5.504126 O        token[0].text()[-3:]:day\n",
            "4.715026 I-MISC   pos_tag[0]:$\n",
            "4.225288 I-PER    token[0].text()[:2]:Mc\n",
            "3.850501 O        token[0].text()[:3]:Sep\n",
            "3.712716 I-LOC    token[-1].text():Colo\n",
            "3.712716 I-LOC    token[-1].lower():colo\n",
            "3.704106 O        token[0].text()[:3]:Aug\n",
            "3.622298 I-ORG    token[0].text()[-2:]:EC\n",
            "3.610589 O        bias\n",
            "3.440807 O        token[0].text()[:2]:W1\n",
            "\n",
            "Top negative:\n",
            "-2.604805 I-ORG    token[0].isdigit()\n",
            "-2.621656 B-LOC    bias\n",
            "-2.621730 O        token[0].text()[-2:]:TO\n",
            "-2.642925 O        pos_tag[0]:NNPS\n",
            "-2.656006 B-MISC   bias\n",
            "-2.771128 O        token[0].text()[:2]:Ro\n",
            "-2.938220 O        token[-1].lower():moody\n",
            "-2.947140 O        token[0].text()[-2:]:AN\n",
            "-3.125447 I-LOC    token[1].isdigit()\n",
            "-3.216320 O        token[0].istitle()\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "---\n",
        "\n",
        "## Hyperparameter optimization\n",
        "\n",
        "(This part of the notebook is copied directly from the [notebook for CoNLL 2002 Spanish tagging](https://github.com/TeamHG-Memex/sklearn-crfsuite/blob/master/docs/CoNLL2002.ipynb))"
      ],
      "metadata": {
        "id": "bXQrG3w4To7a"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "%%time\n",
        "\n",
        "# define fixed parameters and parameters to search\n",
        "crf = sklearn_crfsuite.CRF(\n",
        "    algorithm='lbfgs', \n",
        "    max_iterations=100, \n",
        "    all_possible_transitions=True\n",
        ")\n",
        "\n",
        "param_space = {\n",
        "    'c1': scipy.stats.expon(scale=0.5),\n",
        "    'c2': scipy.stats.expon(scale=0.05),\n",
        "}\n",
        "\n",
        "# use the same metric for evaluation\n",
        "f1_scorer = make_scorer(\n",
        "    metrics.flat_f1_score,\n",
        "    average='weighted',\n",
        "    labels=labels\n",
        ")\n",
        "\n",
        "# search\n",
        "rs = RandomizedSearchCV(\n",
        "    crf,\n",
        "    param_space, \n",
        "    cv=3, \n",
        "    verbose=1, \n",
        "    n_jobs=-1, \n",
        "    n_iter=50, \n",
        "    scoring=f1_scorer\n",
        ")\n",
        "\n",
        "rs.fit(X_train, y_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 173
        },
        "id": "TnAbdJBwTx_j",
        "outputId": "b4186d78-ea3a-4ceb-a5ea-1135ce385f4e"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 3 folds for each of 50 candidates, totalling 150 fits\n",
            "CPU times: user 13min 16s, sys: 1min 33s, total: 14min 50s\n",
            "Wall time: 1h 28min 29s\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomizedSearchCV(cv=3,\n",
              "                   estimator=CRF(algorithm='lbfgs',\n",
              "                                 all_possible_transitions=True,\n",
              "                                 max_iterations=100),\n",
              "                   n_iter=50, n_jobs=-1,\n",
              "                   param_distributions={'c1': <scipy.stats._distn_infrastructure.rv_continuous_frozen object at 0x7f4340460670>,\n",
              "                                        'c2': <scipy.stats._distn_infrastructure.rv_continuous_frozen object at 0x7f43403e7580>},\n",
              "                   scoring=make_scorer(flat_f1_score, average=weighted, labels=['I-ORG', 'I-MISC', 'I-PER', 'I-LOC', 'B-LOC', 'B-MISC', 'B-ORG']),\n",
              "                   verbose=1)"
            ],
            "text/html": [
              "<style>#sk-container-id-2 {color: black;background-color: white;}#sk-container-id-2 pre{padding: 0;}#sk-container-id-2 div.sk-toggleable {background-color: white;}#sk-container-id-2 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-2 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-2 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-2 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-2 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-2 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-2 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-2 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-2 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-2 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-2 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-2 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-2 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-2 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-2 div.sk-item {position: relative;z-index: 1;}#sk-container-id-2 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-2 div.sk-item::before, #sk-container-id-2 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-2 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-2 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-2 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-2 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-2 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-2 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-2 div.sk-label-container {text-align: center;}#sk-container-id-2 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-2 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-2\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>RandomizedSearchCV(cv=3,\n",
              "                   estimator=CRF(algorithm=&#x27;lbfgs&#x27;,\n",
              "                                 all_possible_transitions=True,\n",
              "                                 max_iterations=100),\n",
              "                   n_iter=50, n_jobs=-1,\n",
              "                   param_distributions={&#x27;c1&#x27;: &lt;scipy.stats._distn_infrastructure.rv_continuous_frozen object at 0x7f4340460670&gt;,\n",
              "                                        &#x27;c2&#x27;: &lt;scipy.stats._distn_infrastructure.rv_continuous_frozen object at 0x7f43403e7580&gt;},\n",
              "                   scoring=make_scorer(flat_f1_score, average=weighted, labels=[&#x27;I-ORG&#x27;, &#x27;I-MISC&#x27;, &#x27;I-PER&#x27;, &#x27;I-LOC&#x27;, &#x27;B-LOC&#x27;, &#x27;B-MISC&#x27;, &#x27;B-ORG&#x27;]),\n",
              "                   verbose=1)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomizedSearchCV</label><div class=\"sk-toggleable__content\"><pre>RandomizedSearchCV(cv=3,\n",
              "                   estimator=CRF(algorithm=&#x27;lbfgs&#x27;,\n",
              "                                 all_possible_transitions=True,\n",
              "                                 max_iterations=100),\n",
              "                   n_iter=50, n_jobs=-1,\n",
              "                   param_distributions={&#x27;c1&#x27;: &lt;scipy.stats._distn_infrastructure.rv_continuous_frozen object at 0x7f4340460670&gt;,\n",
              "                                        &#x27;c2&#x27;: &lt;scipy.stats._distn_infrastructure.rv_continuous_frozen object at 0x7f43403e7580&gt;},\n",
              "                   scoring=make_scorer(flat_f1_score, average=weighted, labels=[&#x27;I-ORG&#x27;, &#x27;I-MISC&#x27;, &#x27;I-PER&#x27;, &#x27;I-LOC&#x27;, &#x27;B-LOC&#x27;, &#x27;B-MISC&#x27;, &#x27;B-ORG&#x27;]),\n",
              "                   verbose=1)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: CRF</label><div class=\"sk-toggleable__content\"><pre>CRF(algorithm=&#x27;lbfgs&#x27;, all_possible_transitions=True, max_iterations=100)</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-4\" type=\"checkbox\" ><label for=\"sk-estimator-id-4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">CRF</label><div class=\"sk-toggleable__content\"><pre>CRF(algorithm=&#x27;lbfgs&#x27;, all_possible_transitions=True, max_iterations=100)</pre></div></div></div></div></div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print('best params:', rs.best_params_)\n",
        "print('best CV score:', rs.best_score_)\n",
        "print('model size: {:0.2f}M'.format(rs.best_estimator_.size_ / 1000000))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "a9Ndwu9dT_UI",
        "outputId": "b7f0a201-c454-4a14-c46a-369dd5d9dd12"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "best params: {'c1': 0.010029012775409568, 'c2': 0.008364141984784803}\n",
            "best CV score: 0.8646890591997062\n",
            "model size: 4.10M\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Visualize parameter space"
      ],
      "metadata": {
        "id": "SQoralW_W_kA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "rs.cv_results_"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ck4fV97NoR2B",
        "outputId": "4ecbb7b2-c8d2-47ab-c616-0676a93f77df"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mean_fit_time': array([68.2450858 , 67.80299457, 68.41943558, 68.43861198, 67.9520479 ,\n",
              "        66.79955355, 63.85548536, 61.96898826, 62.19432608, 58.81479526,\n",
              "        59.0412159 , 58.96159983, 60.95325788, 60.7336607 , 59.20995005,\n",
              "        60.07738511, 58.86902014, 61.16866191, 59.57398629, 60.25752775,\n",
              "        58.63062215, 61.52501186, 60.47810594, 59.96968691, 63.812977  ,\n",
              "        59.61794146, 60.51773938, 61.33904624, 64.46386806, 64.85921327,\n",
              "        68.82925518, 69.18700345, 69.29359929, 73.35335294, 62.35595926,\n",
              "        62.5009702 , 61.34446232, 62.37624351, 61.65433526, 61.7156326 ,\n",
              "        60.51364295, 59.38689558, 59.10955691, 59.51269635, 62.07210716,\n",
              "        62.83765093, 60.63622244, 62.09839781, 61.06899865, 55.96699484]),\n",
              " 'std_fit_time': array([3.47226247, 6.36591669, 3.04548495, 7.11084989, 5.11464527,\n",
              "        6.50226437, 7.27822239, 4.91877775, 4.80440044, 4.58468373,\n",
              "        3.79630157, 5.1388931 , 3.88309341, 5.81784906, 1.90163441,\n",
              "        5.5680909 , 5.17460007, 4.47609022, 4.96891271, 3.64276891,\n",
              "        4.40122143, 4.89488103, 4.93384893, 4.72832494, 5.37046871,\n",
              "        4.71575465, 2.94143439, 5.16059084, 5.78447731, 6.16638654,\n",
              "        4.58321893, 4.55471807, 5.15551064, 1.33690492, 4.2444216 ,\n",
              "        5.05515998, 4.78845212, 5.01097471, 4.03490345, 4.19121144,\n",
              "        4.44493159, 3.21035481, 3.65482681, 4.79720607, 2.57258681,\n",
              "        6.1475649 , 3.72300647, 3.80503689, 3.13154699, 8.56589657]),\n",
              " 'mean_score_time': array([3.54322306, 3.12237708, 3.23290563, 3.9350214 , 2.39536277,\n",
              "        2.2575891 , 3.40972424, 2.19823623, 2.46692562, 2.57407554,\n",
              "        2.16996352, 2.9058032 , 2.41259329, 2.33553267, 2.27062861,\n",
              "        2.12725623, 2.14260848, 2.6373659 , 2.52911313, 2.16769425,\n",
              "        2.19465939, 2.21259729, 2.1744101 , 2.59490267, 2.47253442,\n",
              "        2.14710037, 2.34168092, 3.16051825, 2.79744935, 3.62998748,\n",
              "        2.38463386, 2.45509505, 3.39598314, 3.0450534 , 2.20330008,\n",
              "        2.19056042, 2.69968677, 2.60141738, 2.11506756, 2.1734906 ,\n",
              "        2.27479529, 2.62236619, 2.28913069, 2.63649519, 3.43927526,\n",
              "        2.62120779, 2.18189669, 2.25298603, 3.38777884, 2.39036989]),\n",
              " 'std_score_time': array([1.08641352, 1.19603756, 1.26049154, 0.21820602, 0.22252722,\n",
              "        0.21271251, 1.50423473, 0.260714  , 0.26241962, 0.5319948 ,\n",
              "        0.16120703, 1.25086742, 0.35081915, 0.37369655, 0.11987846,\n",
              "        0.21074723, 0.18343593, 0.7603822 , 0.44953871, 0.21447788,\n",
              "        0.18451706, 0.22495368, 0.20157367, 0.16374709, 0.4680487 ,\n",
              "        0.1643176 , 0.21715275, 1.1977555 , 0.46746919, 1.14932539,\n",
              "        0.15697807, 0.37115279, 0.84195646, 0.45270159, 0.21628914,\n",
              "        0.18146198, 0.71334105, 0.84661449, 0.15624292, 0.16434584,\n",
              "        0.1966094 , 0.42364199, 0.19090733, 0.68874477, 0.45174242,\n",
              "        0.8226777 , 0.13816156, 0.15237572, 0.17316712, 0.46824631]),\n",
              " 'param_c1': masked_array(data=[2.2574679661881802, 1.3648786404510316,\n",
              "                    0.7647241406301244, 0.24101460722086968,\n",
              "                    1.5296037625984793, 0.12719627954179738,\n",
              "                    0.24975293613449975, 0.19810143662739696,\n",
              "                    0.18400674764528596, 0.6617029403991295,\n",
              "                    0.11873514204807632, 0.2495728009388029,\n",
              "                    0.5696561382524785, 0.31935946471135473,\n",
              "                    0.7106630458085398, 1.1101756649358394,\n",
              "                    1.150385606074146, 0.14595171090924094,\n",
              "                    0.4293801905926132, 0.4001725183218407,\n",
              "                    1.0010141778914197, 0.1962292024525297,\n",
              "                    0.48966501492747044, 1.0827774056015937,\n",
              "                    0.8260387202895122, 0.6149829462127447,\n",
              "                    0.16071831568219952, 0.006149609729897203,\n",
              "                    0.5589312462385683, 0.7395629332118048,\n",
              "                    0.20789636742112239, 0.2142737780157616,\n",
              "                    0.2554535877161087, 2.9070635488442744,\n",
              "                    0.053219517013360654, 0.2128088997789274,\n",
              "                    0.010029012775409568, 0.19417579965926926,\n",
              "                    1.740477560779506, 0.3573142257671651,\n",
              "                    0.22459567734267039, 0.0030696561422002004,\n",
              "                    0.06731154264688495, 1.9107891643208155,\n",
              "                    0.2906448716301685, 0.4229230514170532,\n",
              "                    0.6015609752618595, 0.18621487572957957,\n",
              "                    0.7148341871458032, 0.14353101345087982],\n",
              "              mask=[False, False, False, False, False, False, False, False,\n",
              "                    False, False, False, False, False, False, False, False,\n",
              "                    False, False, False, False, False, False, False, False,\n",
              "                    False, False, False, False, False, False, False, False,\n",
              "                    False, False, False, False, False, False, False, False,\n",
              "                    False, False, False, False, False, False, False, False,\n",
              "                    False, False],\n",
              "        fill_value='?',\n",
              "             dtype=object),\n",
              " 'param_c2': masked_array(data=[0.061385119095616725, 0.09697226441348333,\n",
              "                    0.0019789007101005144, 0.056036583941296896,\n",
              "                    0.01788862586362565, 0.016227952392521217,\n",
              "                    0.014757756194454836, 0.020111027564127974,\n",
              "                    0.04441147966953743, 0.00226951810359071,\n",
              "                    0.1263046659095782, 0.01858228136043434,\n",
              "                    0.01923245557556287, 0.07918509713240175,\n",
              "                    0.005931095603368696, 0.042258818719288355,\n",
              "                    0.009573750743215795, 0.03933547557091854,\n",
              "                    0.10473734041183858, 0.01817838827501114,\n",
              "                    0.05236941525843989, 0.0011682417830205308,\n",
              "                    0.011494678029997367, 0.12244550455458938,\n",
              "                    0.048335489220805675, 0.004585144232404345,\n",
              "                    0.005023636477634503, 0.01828912089014731,\n",
              "                    0.038148327226119744, 0.11090941064643965,\n",
              "                    0.0026048296989580957, 0.061752587407717933,\n",
              "                    0.09107467092018119, 0.032809138875323766,\n",
              "                    0.0013773226249599573, 0.03544305104020245,\n",
              "                    0.008364141984784803, 0.0036664570507661583,\n",
              "                    0.041944478154371044, 0.01871727802738601,\n",
              "                    0.08125931417620283, 0.08771800023776266,\n",
              "                    0.004849511006991504, 0.012033602917398351,\n",
              "                    0.20770192252749087, 0.05642747577702593,\n",
              "                    0.0056158444850837284, 0.03555443134486672,\n",
              "                    0.03144203986020496, 0.02596416493772573],\n",
              "              mask=[False, False, False, False, False, False, False, False,\n",
              "                    False, False, False, False, False, False, False, False,\n",
              "                    False, False, False, False, False, False, False, False,\n",
              "                    False, False, False, False, False, False, False, False,\n",
              "                    False, False, False, False, False, False, False, False,\n",
              "                    False, False, False, False, False, False, False, False,\n",
              "                    False, False],\n",
              "        fill_value='?',\n",
              "             dtype=object),\n",
              " 'params': [{'c1': 2.2574679661881802, 'c2': 0.061385119095616725},\n",
              "  {'c1': 1.3648786404510316, 'c2': 0.09697226441348333},\n",
              "  {'c1': 0.7647241406301244, 'c2': 0.0019789007101005144},\n",
              "  {'c1': 0.24101460722086968, 'c2': 0.056036583941296896},\n",
              "  {'c1': 1.5296037625984793, 'c2': 0.01788862586362565},\n",
              "  {'c1': 0.12719627954179738, 'c2': 0.016227952392521217},\n",
              "  {'c1': 0.24975293613449975, 'c2': 0.014757756194454836},\n",
              "  {'c1': 0.19810143662739696, 'c2': 0.020111027564127974},\n",
              "  {'c1': 0.18400674764528596, 'c2': 0.04441147966953743},\n",
              "  {'c1': 0.6617029403991295, 'c2': 0.00226951810359071},\n",
              "  {'c1': 0.11873514204807632, 'c2': 0.1263046659095782},\n",
              "  {'c1': 0.2495728009388029, 'c2': 0.01858228136043434},\n",
              "  {'c1': 0.5696561382524785, 'c2': 0.01923245557556287},\n",
              "  {'c1': 0.31935946471135473, 'c2': 0.07918509713240175},\n",
              "  {'c1': 0.7106630458085398, 'c2': 0.005931095603368696},\n",
              "  {'c1': 1.1101756649358394, 'c2': 0.042258818719288355},\n",
              "  {'c1': 1.150385606074146, 'c2': 0.009573750743215795},\n",
              "  {'c1': 0.14595171090924094, 'c2': 0.03933547557091854},\n",
              "  {'c1': 0.4293801905926132, 'c2': 0.10473734041183858},\n",
              "  {'c1': 0.4001725183218407, 'c2': 0.01817838827501114},\n",
              "  {'c1': 1.0010141778914197, 'c2': 0.05236941525843989},\n",
              "  {'c1': 0.1962292024525297, 'c2': 0.0011682417830205308},\n",
              "  {'c1': 0.48966501492747044, 'c2': 0.011494678029997367},\n",
              "  {'c1': 1.0827774056015937, 'c2': 0.12244550455458938},\n",
              "  {'c1': 0.8260387202895122, 'c2': 0.048335489220805675},\n",
              "  {'c1': 0.6149829462127447, 'c2': 0.004585144232404345},\n",
              "  {'c1': 0.16071831568219952, 'c2': 0.005023636477634503},\n",
              "  {'c1': 0.006149609729897203, 'c2': 0.01828912089014731},\n",
              "  {'c1': 0.5589312462385683, 'c2': 0.038148327226119744},\n",
              "  {'c1': 0.7395629332118048, 'c2': 0.11090941064643965},\n",
              "  {'c1': 0.20789636742112239, 'c2': 0.0026048296989580957},\n",
              "  {'c1': 0.2142737780157616, 'c2': 0.061752587407717933},\n",
              "  {'c1': 0.2554535877161087, 'c2': 0.09107467092018119},\n",
              "  {'c1': 2.9070635488442744, 'c2': 0.032809138875323766},\n",
              "  {'c1': 0.053219517013360654, 'c2': 0.0013773226249599573},\n",
              "  {'c1': 0.2128088997789274, 'c2': 0.03544305104020245},\n",
              "  {'c1': 0.010029012775409568, 'c2': 0.008364141984784803},\n",
              "  {'c1': 0.19417579965926926, 'c2': 0.0036664570507661583},\n",
              "  {'c1': 1.740477560779506, 'c2': 0.041944478154371044},\n",
              "  {'c1': 0.3573142257671651, 'c2': 0.01871727802738601},\n",
              "  {'c1': 0.22459567734267039, 'c2': 0.08125931417620283},\n",
              "  {'c1': 0.0030696561422002004, 'c2': 0.08771800023776266},\n",
              "  {'c1': 0.06731154264688495, 'c2': 0.004849511006991504},\n",
              "  {'c1': 1.9107891643208155, 'c2': 0.012033602917398351},\n",
              "  {'c1': 0.2906448716301685, 'c2': 0.20770192252749087},\n",
              "  {'c1': 0.4229230514170532, 'c2': 0.05642747577702593},\n",
              "  {'c1': 0.6015609752618595, 'c2': 0.0056158444850837284},\n",
              "  {'c1': 0.18621487572957957, 'c2': 0.03555443134486672},\n",
              "  {'c1': 0.7148341871458032, 'c2': 0.03144203986020496},\n",
              "  {'c1': 0.14353101345087982, 'c2': 0.02596416493772573}],\n",
              " 'split0_test_score': array([0.81556521, 0.83386929, 0.84698461, 0.85804263, 0.82940574,\n",
              "        0.86020473, 0.85742222, 0.8586738 , 0.85947212, 0.85100511,\n",
              "        0.85933003, 0.85867267, 0.85236384, 0.85481424, 0.8474636 ,\n",
              "        0.8374165 , 0.83819327, 0.860393  , 0.85311587, 0.85621221,\n",
              "        0.84028163, 0.85794769, 0.8539386 , 0.83771711, 0.84557405,\n",
              "        0.85177572, 0.86129993, 0.86930767, 0.85171798, 0.84660046,\n",
              "        0.85940151, 0.85743891, 0.85592755, 0.80557045, 0.86396879,\n",
              "        0.85778007, 0.86890273, 0.85928257, 0.82606807, 0.8560721 ,\n",
              "        0.85713567, 0.86912863, 0.86456699, 0.82186594, 0.85542234,\n",
              "        0.85580961, 0.85203848, 0.85844854, 0.84756098, 0.86011713]),\n",
              " 'split1_test_score': array([0.8348879 , 0.8503223 , 0.86005505, 0.86700419, 0.84820539,\n",
              "        0.87328524, 0.86991604, 0.86977889, 0.86891176, 0.86237728,\n",
              "        0.87076254, 0.8694699 , 0.86168973, 0.86520532, 0.8613971 ,\n",
              "        0.85302806, 0.85227078, 0.87013467, 0.86346725, 0.86621097,\n",
              "        0.85739839, 0.86941099, 0.86094001, 0.8536422 , 0.85721599,\n",
              "        0.86213498, 0.87022312, 0.88059314, 0.86443305, 0.85944638,\n",
              "        0.87069064, 0.86736933, 0.8671117 , 0.82689845, 0.87387215,\n",
              "        0.86710392, 0.88052134, 0.86923467, 0.84584812, 0.86575879,\n",
              "        0.86803834, 0.87799285, 0.87642692, 0.8431896 , 0.8651633 ,\n",
              "        0.86499839, 0.86374128, 0.86986408, 0.85954582, 0.87189772]),\n",
              " 'split2_test_score': array([0.79148376, 0.80848393, 0.8218637 , 0.83512468, 0.80538448,\n",
              "        0.83853544, 0.83492784, 0.83741762, 0.83722654, 0.82460287,\n",
              "        0.8374066 , 0.8358567 , 0.82848592, 0.83212948, 0.82364626,\n",
              "        0.81447891, 0.81479906, 0.83798665, 0.8294059 , 0.83175654,\n",
              "        0.81859552, 0.83890034, 0.83113249, 0.81481741, 0.82202065,\n",
              "        0.82733282, 0.83892898, 0.8434845 , 0.82807923, 0.82250729,\n",
              "        0.83641897, 0.83519641, 0.83525867, 0.77928002, 0.84198952,\n",
              "        0.83619573, 0.84464311, 0.83785579, 0.80107007, 0.83206197,\n",
              "        0.83386531, 0.84357333, 0.84115271, 0.79945847, 0.82965589,\n",
              "        0.8305439 , 0.82697599, 0.83780771, 0.82504299, 0.83796623]),\n",
              " 'mean_test_score': array([0.81397895, 0.83089184, 0.84296779, 0.8533905 , 0.8276652 ,\n",
              "        0.8573418 , 0.8540887 , 0.8552901 , 0.85520348, 0.84599509,\n",
              "        0.85583305, 0.85466642, 0.84751316, 0.85071635, 0.84416899,\n",
              "        0.83497449, 0.83508771, 0.85617144, 0.84866301, 0.85139324,\n",
              "        0.83875851, 0.85541967, 0.84867037, 0.83539224, 0.84160356,\n",
              "        0.84708117, 0.85681734, 0.86446177, 0.84807675, 0.84285138,\n",
              "        0.85550371, 0.85333488, 0.85276597, 0.80391631, 0.85994349,\n",
              "        0.85369324, 0.86468906, 0.85545767, 0.82432875, 0.85129762,\n",
              "        0.85301311, 0.86356493, 0.86071554, 0.82150467, 0.85008051,\n",
              "        0.85045063, 0.84758525, 0.85537344, 0.84404993, 0.85666036]),\n",
              " 'std_test_score': array([0.01775513, 0.01720971, 0.01584815, 0.01342405, 0.01752484,\n",
              "        0.01433026, 0.01447706, 0.01342634, 0.01328293, 0.01582301,\n",
              "        0.01384019, 0.01401188, 0.01398261, 0.01381056, 0.01558679,\n",
              "        0.01583207, 0.01545457, 0.01345957, 0.01425751, 0.01447282,\n",
              "        0.01587777, 0.01258354, 0.01272629, 0.01593518, 0.01464016,\n",
              "        0.01459056, 0.01316311, 0.01553222, 0.01506307, 0.01531156,\n",
              "        0.01426023, 0.01345131, 0.01319471, 0.0194753 , 0.01332361,\n",
              "        0.0129449 , 0.0149472 , 0.01309277, 0.01832189, 0.01416488,\n",
              "        0.01425238, 0.01459205, 0.01465589, 0.01785498, 0.01497989,\n",
              "        0.01456747, 0.01533613, 0.01326637, 0.01430284, 0.01406647]),\n",
              " 'rank_test_score': array([49, 45, 38, 20, 46,  6, 18, 15, 16, 35, 10, 17, 33, 26, 36, 44, 43,\n",
              "         9, 30, 24, 41, 13, 29, 42, 40, 34,  7,  2, 31, 39, 11, 21, 23, 50,\n",
              "         5, 19,  1, 12, 47, 25, 22,  3,  4, 48, 28, 27, 32, 14, 37,  8],\n",
              "       dtype=int32)}"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%matplotlib inline\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "plt.style.use('ggplot')"
      ],
      "metadata": {
        "id": "xkuvRt7LovDm"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "_x = [s['c1'] for s in rs.cv_results_['params']]\n",
        "_y = [s['c2'] for s in rs.cv_results_['params']]\n",
        "_c = [s for s in rs.cv_results_['mean_test_score']]\n",
        "\n",
        "fig = plt.figure()\n",
        "fig.set_size_inches(12, 12)\n",
        "ax = plt.gca()\n",
        "ax.set_yscale('log')\n",
        "ax.set_xscale('log')\n",
        "ax.set_xlabel('C1')\n",
        "ax.set_ylabel('C2')\n",
        "ax.set_title(\"Randomized Hyperparameter Search CV Results (min={:0.3}, max={:0.3})\".format(\n",
        "    min(_c), max(_c)\n",
        "))\n",
        "\n",
        "ax.scatter(_x, _y, c=_c, s=60, alpha=0.9, edgecolors=[0,0,0])\n",
        "\n",
        "print(\"Dark blue => {:0.4}, dark red => {:0.4}\".format(min(_c), max(_c)))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "h_Ln8o68W5Z7",
        "outputId": "ace025a8-7da1-408d-b7f0-53bf4eb8a972"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Dark blue => 0.8039, dark red => 0.8647\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x1200 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "crf = rs.best_estimator_\n",
        "\n",
        "y_pred = crf.predict(X_devel)\n",
        "\n",
        "print(metrics.flat_classification_report(\n",
        "    y_devel, y_pred, labels=sorted_labels, digits=3\n",
        "))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "p30qN0K-XDtx",
        "outputId": "0742477a-a85b-4794-8623-d76749b1405d"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "       B-LOC      0.000     0.000     0.000         0\n",
            "       I-LOC      0.915     0.884     0.900      2094\n",
            "      B-MISC      0.250     0.250     0.250         4\n",
            "      I-MISC      0.910     0.828     0.867      1264\n",
            "       B-ORG      0.000     0.000     0.000         0\n",
            "       I-ORG      0.874     0.833     0.853      2092\n",
            "       I-PER      0.942     0.925     0.934      3149\n",
            "\n",
            "   micro avg      0.914     0.878     0.896      8603\n",
            "   macro avg      0.556     0.532     0.543      8603\n",
            "weighted avg      0.914     0.878     0.896      8603\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.9/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        }
      ]
    }
  ]
}