ML

Naive machine learning algorithm implementations in c++. I built these while taking Andrew Ng's machine learning course on Coursera.

Models

• Linear Regression
• Logistic Regression (Binary and Multiclass)
• Neural Network
• SVM (https://github.com/mazefeng/svm/blob/master/svm_solver.cpp)

How to run

To run the code examples you will need to have Eigen installed.

Linear Regression

Navigate to the LinearRegression directory and run:

make
./mlinreg n < data/mdata1.txt

This will train a multivariate linear regression model on some data. The n flag tells the program to normalize the input data.

Other examples are found in the data subdirectory.

Logistic Regression

Navigate to the LogisticRegression directory and run:

make
./logreg n < data/data2.txt

This will train a logistic regression model on some data. The n flag tells the program to normalize the input data.

Other examples are found in the data subdirectory.

Neural Network

Navigate to the NeuralNetwork directory and run:

make
./nn d data/digits.txt

This will train a neural network to predict a digit from a 20x20 input image.

Support Vector Machine

Navigate to the SVM directory and run:

make
./svm < data/data2.txt

This will train a SVM model on some data.

Other examples are found in the data subdirectory.