Create spam classification tutorial

.travis.yml

@@ -1,5 +1,5 @@
 sudo: required
-dist: xenial
+dist: focal
 language: cpp
 
 before_install:
@@ -46,7 +46,7 @@ before_install:
   - cd mlpack
   - mkdir build
   - cd build
-  - cmake -DCMAKE_INSTALL_PREFIX=/usr -DUSE_OPENMP=OFF -DBUILD_CLI_EXECUTABLES=OFF -DBUILD_JULIA_BINDINGS=OFF -DBUILD_PYTHON_BINDINGS=OFF -DBUILD_MARKDOWN_BINDINGS=OFF -DBUILD_R_BINDINGS=OFF -DBUILD_TESTS=OFF ..
+  - cmake -DCMAKE_INSTALL_PREFIX=/usr -DUSE_OPENMP=ON -DBUILD_CLI_EXECUTABLES=ON -DBUILD_JULIA_BINDINGS=OFF -DBUILD_PYTHON_BINDINGS=OFF -DBUILD_MARKDOWN_BINDINGS=OFF -DBUILD_R_BINDINGS=OFF -DBUILD_TESTS=OFF ..
   - make -j2
   - sudo make install
   - cd ../
@@ -58,32 +58,38 @@ install:
   # Download datasets.
   - pwd
   - pip install tqdm
-  - cd tools/
-  - ./download_data_set.py
-  - cd ../
-  - ls data/
+    - cd tools/
+    - ./download_data_set.py
+    - cd ../
+    - ls data/
 
 script:
   # Finally, build all the examples.
   - pwd
-  - ls
-  - ls data/
-  - |
+    - ls
+    - ls data/
+    - |
     for f in */Makefile; do
       dir=`dirname $f`;
-
+    
       # TODO: this takes too long right now.
       if [ "$dir" == "mnist_cnn" ];
       then
         continue;
       fi
-
+    
       # TODO: the dataset cannot be loaded.	
       if [ "$dir" == "mnist_vae_cnn" ];	
       then	
         continue;	
       fi	
-
+
+      # TODO: the dataset cannot be loaded.	
+      if [ "$dir" == "spam" ];	
+      then	
+        continue;	
+      fi
+
       cd $dir;
       target_name=`grep 'TARGET := ' Makefile | awk -F ' := ' '{ print $2 }'`;
       echo "Make target $target_name in directory $dir.";
@@ -98,17 +104,16 @@ script:
       fi
       cd ../;
     done
-  # Print any failures.
-  - ls
-  - |
+      # Print any failures.
+    - ls
+    - |
     if [ -f faillog ]; then
       echo "Some examples failed!";
       echo "";
       cat faillog;
       exit 1;
     fi
 
-
 notifications:
   email:
     - [email protected]

README.md

@@ -61,6 +61,9 @@ description (just a little bit more than the title):
  - `lstm_stock_prediction`: predict Google's historical stock price (daily high
    _and_ low) using an LSTM-based recurrent neural network
 
+ - `spam`: predict whether a mobile phone text message in Indonesian is spam 
+    or not using logistic regression
+
  - `mnist_batch_norm`: use batch normalization in a simple feedforward neural
    network to recognize the MNIST digits