package classification;

import helperClasses.DataPoint;

import helperClasses.Serialization;

import java.util.ArrayList;

class NaiveBayes {

public ArrayList<DataPoint> trainer = new ArrayList<DataPoint>();

public void importTrainingData(){

ArrayList<DataPoint> t = null;

Serialization S = new Serialization();

trainer = S.deserialize(t, "nbcTrainingData.ser");

}

public void updateTrainingData(DataPoint newData){

//Used to create the training data (".ser" file)

/*trainer.add(new DataPoint("Sunny", "Hot", "H", "W", "NO"));

trainer.add(new DataPoint("Sunny", "Hot", "H", "S", "NO"));

trainer.add(new DataPoint("Overcast", "Hot", "H", "W", "YES"));

trainer.add(new DataPoint("Rain", "Mild", "H", "W", "YES"));

trainer.add(new DataPoint("Rain", "Cool", "N", "W", "YES"));

trainer.add(new DataPoint("Rain", "Cool", "N", "S", "NO"));

trainer.add(new DataPoint("Overcast", "Cool", "N", "S", "YES"));

trainer.add(new DataPoint("Sunny", "Mild", "H", "W", "NO"));

trainer.add(new DataPoint("Sunny", "Cool", "N", "W", "YES"));

trainer.add(new DataPoint("Rain", "Mild", "N", "W", "YES"));

trainer.add(new DataPoint("Sunny", "Mild", "N", "S", "YES"));

trainer.add(new DataPoint("Overcast", "Mild", "H", "S", "YES"));

trainer.add(new DataPoint("Overcast", "Hot", "N", "W", "YES"));

trainer.add(new DataPoint("Rain", "Mild", "H", "S", "NO"));*/

trainer.add(newData);

Serialization S = new Serialization();

S.serialize(trainer, "nbcTrainingData.ser");

}

public void dispayTrainingData(){

ArrayList<DataPoint> T = null;

Serialization S = new Serialization();

T = S.deserialize(T, "nbcTrainingData.ser");

System.out.println("Outlook\t\tTemperature\tHumidity\tWind\tClass");

for (DataPoint P : T)

System.out.println(P.outlook+"\t\t"+P.temperature+"\t\t"+P.humidity+"\t\t"+P.wind+"\t"+P.C);

}

public void nbcClassify(DataPoint input){

double nYES=0, nNO=0, nOY=0, nTY=0, nHY=0, nWY=0, nON=0, nTN=0, nHN=0, nWN=0;

double total = trainer.size();

for(DataPoint R : trainer){

if(R.C.equals("YES")){

nYES++;

if(R.outlook.equals(input.outlook)) nOY++;

if(R.temperature.equals(input.temperature)) nTY++;

if(R.humidity.equals(input.humidity)) nHY++;

if(R.wind.equals(input.wind)) nWY++;

continue;

}

else{

nNO++;

if(R.outlook.equals(input.outlook)) nON++;

if(R.temperature.equals(input.temperature)) nTN++;

if(R.humidity.equals(input.humidity)) nHN++;

if(R.wind.equals(input.wind)) nWN++;

}

}

double pYES = nYES/total;

double pNO = nNO/total;

double pOY = nOY/nYES;

double pON = nON/nNO;

double pTY = nTY/nYES;

double pTN = nTN/nNO;

double pHY = nHY/nYES;

double pHN = nHN/nNO;

double pWY = nWY/nYES;

double pWN = nWN/nNO;

double inputY = pYES * pOY * pTY * pHY * pWY;

double inputN = pNO * pON * pTN * pHN * pWN;

if(inputY >= inputN) input.C = "YES";

else input.C = "NO";

System.out.println("Class of input : "+input.C);

}

public static void main(String[] args) {

// TODO Auto-generated method stub

NaiveBayes sol = new NaiveBayes();

sol.importTrainingData();

DataPoint input = new DataPoint("Rain", "Hot", "N", "W");

sol.nbcClassify(input);

System.out.println("\n\n\t***********Updated Training Data************\t");

System.out.println("Old Size : "+sol.trainer.size());

sol.updateTrainingData(input);

System.out.println("New Size : "+sol.trainer.size());

sol.dispayTrainingData();

}

}