Naive Bayes

This is a c implementation of the Machine Learning algorithm called Naive Bayes Gaussian. It's been adapted from the python version in: Naive Bayes Classifier from Scratch in Python. This algorithm was developed to be used in the Embbeded System Module at IFCE to be run in an ARM9 platform.

This algorithm is divided into two steps:

• Training, where the dataset will be used to generate the summaries (means and standard deviation) intermediate dataset. This phase happens in a computer.
• Prediction, where the intermediate dataset + testset will be use to generate the predictions. The prediction phase is the one that runs in the ARM9 platform

1) Compiling

To get it up and running in a computer, execute the following steps:

1. Select a dataset by commenting/uncommenting the datasets define section at the defines.h file;
2. Run the training/main.c file to generate the training files; This can also be done by running make -f makefile inside the training folder
3. Run the prediction/main.c file to see the classification report (accuracy, recall, precision etc); This can also be done by running make -f makefile inside the prediction folder
4. Run your program by typing ./main inside the prediction folder

To compile the code for the ARM9 plataform:

• Repeat the steps 1 and 2 in a PC
• Run the command arm-linux-gcc main.c prediction.c extras.c csv.c -o main -lm from within the prediction folder in a PC with the ARM compiler
• In the ARM platform simply run your program by typing ./main from within the prediction folder

2) Project organization

The project was organized based on the STM32F Peripheral Library examples provided by STMicroeletronics followig these guidelines:

• Each .c file has an .h file with the functions headers
• The functions were separated according to their purposes
• The comments can be used to generate automatic documentation

3) Project structure

The project folder structure is defined as follows:

• NaiveBayes - root folder
  • prediction - this folder contains the code used to make the predictions (these are the files that execute the algorithm that runs in the ARM platform)
  • testers - this folder contains the codes for the test code (python with library and original python code)
  • training - this folder contains the code used to generate the summaries and test set
  • datasets - this folder contains all datasets that will be used throughout the execution
    • full - this folder contains the full version of the datasets to be used
    • summaries - this folder contains the summarised knowledge of the classifier (means and standard deviation for each class and attribute)
    • test - this folder contains the test set to be classified