樸素貝葉斯算法---統計學習方法例4.1計算（MATLAB）

個人學習記錄，只進行了書上例題的結果驗證，沒有改寫成函數。

%%%%%%    樸素貝葉斯算法     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

clc
clear

X = {1,'S';1,'M';1,'M';1,'S';1,'S';2,'S';2,'M';2,'M';2,'L';2,'L';3,'L';3,'M';3,'M';3,'L';3,'L'};
Y = [-1,-1,1,1,-1,-1,-1,1,1,1,1,1,1,1,-1]';
x = {2,'S'};  
%%%%%%%  輸入參數X：訓練數據的特徵集   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%          Y：訓練數據的目標集   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%          x：待分類數據的特徵   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%  輸出參數y：待分類數據的類別   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Num = length(Y); %訓練樣本總數
%計算正負類的先驗概率
PositiveIndex = find(Y == 1);
NegativeIndex = find(Y == -1);
PositiveNum = length(PositiveIndex);  %訓練集中正類的個數
NegativeNum = length(NegativeIndex); %訓練集中負類的個數

format rat

PositiveP = PositiveNum/Num; %正類的先驗概率        
NegativeP = NegativeNum/Num; %負類的先驗概率

%計算條件概率，即類別確定的條件下，訓練集中不同特徵取值的概率
PositiveX1 = X(PositiveIndex,1); %正類對應的特徵1
PositiveX2 = X(PositiveIndex,2); %正類對應的特徵2
NegativeX1 = X(NegativeIndex,1); %負類對應的特徵1
NegativeX2 = X(NegativeIndex,2); %負類對應的特徵2  

P_PositiveX11 = length(find(cell2mat(PositiveX1)==1))/PositiveNum;
P_PositiveX12 = length(find(cell2mat(PositiveX1)==2))/PositiveNum;
P_PositiveX13 = length(find(cell2mat(PositiveX1)==3))/PositiveNum;
P_PositiveX21 = length(find(cell2mat(PositiveX2)=='S'))/PositiveNum;
P_PositiveX22 = length(find(cell2mat(PositiveX2)=='M'))/PositiveNum; 
P_PositiveX23 = length(find(cell2mat(PositiveX2)=='L'))/PositiveNum;
P_NegativeX11 = length(find(cell2mat(NegativeX1)==1))/NegativeNum; 
P_NegativeX12 = length(find(cell2mat(NegativeX1)==2))/NegativeNum;  
P_NegativeX13 = length(find(cell2mat(NegativeX1)==3))/NegativeNum;
P_NegativeX21 = length(find(cell2mat(NegativeX2)=='S'))/NegativeNum;  
P_NegativeX22 = length(find(cell2mat(NegativeX2)=='M'))/NegativeNum;  
P_NegativeX23 = length(find(cell2mat(NegativeX2)=='L'))/NegativeNum;

%待分類實例x的後驗概率
P_Positive_x = [PositiveP*P_PositiveX12*P_PositiveX21,1];
P_Negative_x = [NegativeP*P_NegativeX12*P_NegativeX21,-1];

%選擇後驗概率最大的類別作爲實例x的類別
P_x = [P_Positive_x;P_Negative_x];
y = P_x(find(P_x(:,1) == max(P_x(:,1))),2)