HueLogisticRegression/run.m at master · psalv/HueLogisticRegression · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141


%% %%% ========== INIT ========== %%%

clear ; close all ; clc


%%% ========== VARIABLE PARAMETERS ========== %%%

threshold = 0.6;
lambda = 1;


%%% ========== LOADING DATA AND DIVIDING INTO SETS ========== %%%

data = load('testData.txt');
X = data(:, [1, 2]); y = data(:, 3);


%max = 0;
%maxL = -1;
%maxT = -1;
%maxD = -1;

%m = size(y)(1);
%y_train = y([1:floor(m * 0.6)], :);
%y_val = y([floor(m * 0.6) + 1:floor(m * 0.8)], :);
%y_test = y([floor(m * 0.8):end], :);

%for d = 1:10
%  X = mapFeature(X(:,1), X(:,2), d);
%  x_train = X([1:floor(m * 0.6)], :);
%  x_val = X([floor(m * 0.6) + 1:floor(m * 0.8)], :);
%  x_test = X([floor(m * 0.8):end], :);
%  for L = 1:1:50
%    for T = 0.3:0.03:0.9
%      theta = normalEquation(x_train, y_train, L);
%      predicted_y = predict(theta, x_val, threshold);
%      [f, p, r] = fscore(predicted_y, y_val);
%      if f > max
%        max = f;
%        maxL = L;
%        maxT = T;
%        maxD = d;
%      endif
%    end
  %end
%end

%fprintf("\n\n\n\n---------\n\nMax Lambda: %d", maxL);
%fprintf("\nMax F: %d\n", max);


oldX = X;
X = mapFeature(X(:,1), X(:,2), 5);
X(1, :)

m = size(y)(1);

x_train = X([1:floor(m * 0.6)], :);
y_train = y([1:floor(m * 0.6)], :);

x_val = X([floor(m * 0.6) + 1:floor(m * 0.8)], :);
y_val = y([floor(m * 0.6) + 1:floor(m * 0.8)], :);

x_test = X([floor(m * 0.8):end], :);
y_test = y([floor(m * 0.8):end], :);


%%% ========== TESTING OUT VARIOUS LAMBDA AND THRESHOLD ========== %%%

%max = 0;
%maxL = -1;
%maxT = -1;

%for L = 0:0.2:50
%  for T = 0.1:0.02:0.9
%    theta = normalEquation(x_train, y_train, L);
%    predicted_y = predict(theta, x_val, T);
%
    %fprintf("\n\n\nLambda: %d", L);
    %fprintf("\nThreshold: %d\n", T);
%    f = fscore(predicted_y, y_val);
%    if f > max
%      max = f;
%      maxL = L;
%      maxT = T;
%    endif
%  end
%end

%fprintf("\n\n\n\n---------\n\nMax Lambda: %d", maxL);
%fprintf("\nMax Threshold: %d", maxT);
%fprintf("\nMax F: %d\n", max);


%%% ========== PLOTTING THE LEARNING CURVE ========== %%%

%[error_train, error_val] = learningCurve(x_train, y_train, x_val, y_val, lambda, threshold);

%figure(1);
%plot(1:size(x_train)(1), error_train, 1:size(x_train)(1), error_val);
%title('Learning curve for logistic regression')
%legend('Train', 'Cross Validation')
%xlabel('Number of training examples')
%ylabel('Error')
%axis([0 size(x_train)(1) 0 1])


%%% ========== CHECKING PARAMS ========== %%%
theta = normalEquation(x_train, y_train, lambda);
predicted_y = predict(theta, x_val, threshold);

%[oldX([floor(m * 0.6) + 1:floor(m * 0.8)], :)(:, 2) predicted_y y_val]

[f, p, r] = fscore(predicted_y, y_val);

fprintf('Precision: %d\n', p);
fprintf('Recall: %d\n', r);
fprintf('F1 Score: %d\n', f);