Naive Bayes Classifer

Introduction

This project implements the Naive Bayes Classifier in C++. It utilizes the Thread Pool for multi-processing during training and testing.

Other Dependencies

It also makes use of glog, boost library and Json for Modern C++.

Installation

$ configure
$ make
$ make install

Execution Examples

To get help

$ src/naive_bayes -h

To train and test data in Json format

$ src/naive_bayes -J -t train.json -T test.json -l label -f field_map.txt -P8

Explanation

-J: The input file is in Json format.  If is in CSV format, use -C.
-t: The train file in json format.
-T: The test file in json format.
-l: The field name in the json file serves as the label of the entry.
-f: The field map indicates which fields in the json file will
    be used to train and test.  One field name per each line.
-P: Parallel processing for the training and testing.

Comments, questions and pull requests are equally welcome.

Thanks,
-WeeSan <weesan@weesan.com>