randomforest 0.1.6

Random forest regressor and classifier
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

//! Criterions to measure the quality of a node split.
use crate::functions;

/// This trait allows measuring the quality of a node split.
pub trait Criterion: Send + Sync + Clone {
    /// Calculates the quality of a node split (the lower return value, the beter).
    fn calculate<T>(&self, xs: T) -> f64
    where
        T: Iterator<Item = f64> + Clone;
}

/// This trait indicates criterions for regression.
pub trait RegressionCriterion: Criterion {}

/// Mean Squared Error.
#[derive(Debug, Clone)]
pub struct Mse;

impl Criterion for Mse {
    fn calculate<T>(&self, ys: T) -> f64
    where
        T: Iterator<Item = f64> + Clone,
    {
        let n = ys.clone().count() as f64;
        let m = functions::mean(ys.clone());
        ys.map(|x| (x - m).powi(2)).sum::<f64>() / n
    }
}

impl RegressionCriterion for Mse {}

/// This trait indicates criterions for classifications.
pub trait ClassificationCriterion: Criterion {}

/// Gini impurity.
#[derive(Debug, Clone)]
pub struct Gini;

impl Criterion for Gini {
    fn calculate<T>(&self, ys: T) -> f64
    where
        T: Iterator<Item = f64> + Clone,
    {
        let (histogram, n) = functions::histogram(ys);
        1.0 - histogram
            .into_iter()
            .map(|(_, count)| (count as f64 / n as f64).powi(2))
            .sum::<f64>()
    }
}

impl ClassificationCriterion for Gini {}

/// Information entropy.
#[derive(Debug, Clone)]
pub struct Entropy;

impl Criterion for Entropy {
    fn calculate<T>(&self, ys: T) -> f64
    where
        T: Iterator<Item = f64> + Clone,
    {
        let (histogram, n) = functions::histogram(ys);
        histogram
            .into_iter()
            .map(|(_, count)| {
                let p = count as f64 / n as f64;
                -p * p.log2()
            })
            .sum()
    }
}

impl ClassificationCriterion for Entropy {}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn mse_works() {
        assert_eq!(
            Mse.calculate([50.0, 60.0, 70.0, 70.0, 100.0].iter().copied()),
            280.0
        );
    }

    #[test]
    fn gini_works() {
        assert_eq!(
            Gini.calculate([0.0, 1.0, 0.0, 1.0, 1.0, 0.0].iter().copied()),
            0.5
        );
        assert_eq!(
            Gini.calculate([0.0, 0.0, 0.0, 0.0, 0.0, 0.0].iter().copied()),
            0.0
        );
        assert_eq!(
            Gini.calculate([0.0, 1.0, 0.0, 0.0, 0.0, 0.0].iter().copied()),
            0.2777777777777777
        );
    }

    #[test]
    fn entropy_works() {
        assert_eq!(Entropy.calculate([0.0, 1.0, 0.0, 1.0].iter().copied()), 1.0);
        assert_eq!(Entropy.calculate([0.0, 0.0, 0.0, 0.0].iter().copied()), 0.0);
        assert_eq!(
            Entropy.calculate([0.0, 1.0, 0.0, 0.0].iter().copied()),
            0.8112781244591328
        );
    }
}