rust-ml 0.1.5

A collection of machine learning algorithms implemented in pure Rust (personal project for practice).
Documentation 
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
142
143
144
145
146
147
148
149
150
151
152
153

use crate::core::activations::activation::Activation;
use ndarray::{Array, Array1, Array2, Ix1, Ix2};

pub struct Sigmoid;

impl Activation<Ix1> for Sigmoid {
    fn activate(z: &Array1<f64>) -> Array1<f64> {
        1.0 / (1.0 + (-z).exp())
    }

    fn derivative(z: &Array1<f64>) -> Array1<f64> {
        let y_hat = Self::activate(z);
        &y_hat * (1.0 - &y_hat)
    }
}

impl Activation<Ix2> for Sigmoid {
    fn activate(z: &Array2<f64>) -> Array2<f64> {
        1.0 / (1.0 + (-z).exp())
    }

    fn derivative(z: &Array<f64, Ix2>) -> Array<f64, Ix2> {
        let y_hat = Self::activate(z);
        &y_hat * (1.0 - &y_hat)
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use approx::assert_abs_diff_eq;
    use ndarray::{arr1, arr2};
    use ndarray_rand::rand_distr::num_traits::Float;

    #[test]
    fn test_sigmoid_activate_vector() {
        // Test with a vector input
        let input = arr1(&[-2.0, -1.0, 0.0, 1.0, 2.0]);
        let result = Sigmoid::activate(&input);

        // Expected values calculated using sigmoid formula: 1 / (1 + exp(-x))
        let expected = arr1(&[
            0.11920292, // sigmoid(-2.0)
            0.26894142, // sigmoid(-1.0)
            0.5,        // sigmoid(0.0)
            0.73105858, // sigmoid(1.0)
            0.88079708, // sigmoid(2.0)
        ]);

        assert_eq!(result.len(), expected.len());

        // Compare with a small epsilon for floating point precision
        for (a, b) in result.iter().zip(expected.iter()) {
            assert_abs_diff_eq!(a, b, epsilon = 1e-8);
        }
    }

    #[test]
    fn test_sigmoid_derivative_vector() {
        // Test with a vector input
        let input = arr1(&[-2.0, -1.0, 0.0, 1.0, 2.0]);
        let result = Sigmoid::derivative(&input);

        // Expected values calculated using sigmoid derivative formula: sigmoid(x) * (1 - sigmoid(x))
        let sigmoid_values = arr1(&[
            0.11920292, // sigmoid(-2.0)
            0.26894142, // sigmoid(-1.0)
            0.5,        // sigmoid(0.0)
            0.73105858, // sigmoid(1.0)
            0.88079708, // sigmoid(2.0)
        ]);

        let expected = sigmoid_values.mapv(|v| v * (1.0 - v));

        assert_eq!(result.len(), expected.len());

        // Compare with a small epsilon for floating point precision
        for (a, b) in result.iter().zip(expected.iter()) {
            assert_abs_diff_eq!(a, b, epsilon = 1e-8);
        }
    }

    #[test]
    fn test_sigmoid_activate_matrix() {
        // Test with a matrix input
        let input = arr2(&[[-2.0, -1.0, 0.0], [1.0, 2.0, 3.0]]);

        let result = Sigmoid::activate(&input);

        // Expected values calculated using sigmoid formula: 1 / (1 + exp(-x))
        let expected = arr2(&[
            [0.11920292, 0.26894142, 0.5],
            [0.73105858, 0.88079708, 0.95257413],
        ]);

        assert_eq!(result.shape(), expected.shape());

        // Compare with a small epsilon for floating point precision
        for ((i, j), value) in result.indexed_iter() {
            assert_abs_diff_eq!(value, &expected[[i, j]], epsilon = 1e-8);
        }
    }

    #[test]
    fn test_sigmoid_derivative_matrix() {
        // Test with a matrix input
        let input = arr2(&[[-2.0, -1.0, 0.0], [1.0, 2.0, 3.0]]);

        let result = Sigmoid::derivative(&input);

        // Expected values calculated using sigmoid derivative formula: sigmoid(x) * (1 - sigmoid(x))
        let sigmoid_values = arr2(&[
            [0.11920292, 0.26894142, 0.5],
            [0.73105858, 0.88079708, 0.95257413],
        ]);

        let expected = sigmoid_values.mapv(|v| v * (1.0 - v));

        assert_eq!(result.shape(), expected.shape());

        // Compare with a small epsilon for floating point precision
        for ((i, j), value) in result.indexed_iter() {
            assert_abs_diff_eq!(value, &expected[[i, j]], epsilon = 1e-8);
        }
    }

    #[test]
    fn test_sigmoid_properties() {
        // Test that sigmoid(0) = 0.5
        let zero = arr1(&[0.0]);
        assert_abs_diff_eq!(Sigmoid::activate(&zero)[0], 0.5, epsilon = 1e-8);

        // Test symmetry property: sigmoid(-x) = 1 - sigmoid(x)
        let x = arr1(&[1.0, 2.0, 3.0]);
        let neg_x = arr1(&[-1.0, -2.0, -3.0]);

        let sigmoid_x = Sigmoid::activate(&x);
        let sigmoid_neg_x = Sigmoid::activate(&neg_x);

        for (i, value) in sigmoid_neg_x.iter().enumerate() {
            assert_abs_diff_eq!(value, &(1.0 - sigmoid_x[i]), epsilon = 1e-8);
        }

        // Test that derivative is maximum at x = 0
        let points = arr1(&[-2.0, -1.0, -0.5, 0.0, 0.5, 1.0, 2.0]);
        let derivatives = Sigmoid::derivative(&points);

        // Maximum value should be at x = 0 (which is 0.25)
        let max_derivative = derivatives.iter().fold(0.0, |max, &val| max.max(val));
        assert_abs_diff_eq!(max_derivative, 0.25, epsilon = 1e-8);
        assert_abs_diff_eq!(derivatives[3], 0.25, epsilon = 1e-8); // index 3 is x = 0
    }
}