forester 0.0.3

A crate for implementing various flavors of random forests and decision trees.
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
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236

//! Various operations on arrays, slices, and vectors.

use std::iter;
use std::mem;
use std::ops;

use rand::prelude::*;
use rand::distributions::Uniform;

/// Trait that defines a dot product
pub trait Dot<B: ?Sized> {
    type Output;
    fn dot(&self, other: &B) -> Self::Output;
}

impl<T> Dot<[T]> for [T]
    where T: ops::Mul + Clone,
          T::Output: iter::Sum
{
    type Output = T::Output;

    fn dot(&self, other: &[T]) -> Self::Output {
        assert_eq!(self.len(), other.len());
        self.iter().zip(other.iter()).map(|(a, b)| a.clone() * b.clone()).sum()
    }
}

impl<T> Dot<[T]> for Vec<T>
    where T: ops::Mul + Clone,
          T::Output: iter::Sum
{
    type Output = T::Output;

    fn dot(&self, other: &[T]) -> Self::Output {
        assert_eq!(self.len(), other.len());
        self.iter().zip(other.iter()).map(|(a, b)| a.clone() * b.clone()).sum()
    }
}

macro_rules! impl_dot_array_for_array {
    ( $( $size:expr ),* ) => {
        $(
            impl<T> Dot<[T; $size]> for [T; $size]
                where T: ops::Mul + Clone,
                      T::Output: iter::Sum
            {
                type Output = T::Output;
                fn dot(&self, other: &Self) -> Self::Output {
                    self.iter().zip(other.iter()).map(|(a, b)| a.clone() * b.clone()).sum()
                }
            }
        )*
    };
}

impl_dot_array_for_array!{ 1,  2,  3,  4,  5,  6,  7,  8}
impl_dot_array_for_array!{ 9, 10, 11, 12, 13, 14, 15, 16}
impl_dot_array_for_array!{17, 18, 19, 20, 21, 22, 23, 24}
impl_dot_array_for_array!{25, 26, 27, 28, 29, 30, 31, 32}

macro_rules! impl_dot_slice_for_array {
    ( $( $size:expr ),* ) => {
        $(
            impl<T> Dot<[T]> for [T; $size]
                where T: ops::Mul + Clone,
                      T::Output: iter::Sum
            {
                type Output = T::Output;
                fn dot(&self, other: &[T]) -> Self::Output {
                    assert_eq!($size, other.len());
                    self.iter().zip(other.iter()).map(|(a, b)| a.clone() * b.clone()).sum()
                }
            }
        )*
    };
}

impl_dot_slice_for_array!{ 1,  2,  3,  4,  5,  6,  7,  8}
impl_dot_slice_for_array!{ 9, 10, 11, 12, 13, 14, 15, 16}
impl_dot_slice_for_array!{17, 18, 19, 20, 21, 22, 23, 24}
impl_dot_slice_for_array!{25, 26, 27, 28, 29, 30, 31, 32}

macro_rules! impl_dot_array_for_slice {
    ( $( $size:expr ),* ) => {
        $(
            impl<T> Dot<[T; $size]> for [T]
                where T: ops::Mul + Clone,
                      T::Output: iter::Sum
            {
                type Output = T::Output;
                fn dot(&self, other: &[T; $size]) -> Self::Output {
                    assert_eq!($size, self.len());
                    self.iter().zip(other.iter()).map(|(a, b)| a.clone() * b.clone()).sum()
                }
            }
        )*
    };
}

impl_dot_array_for_slice!{ 1,  2,  3,  4,  5,  6,  7,  8}
impl_dot_array_for_slice!{ 9, 10, 11, 12, 13, 14, 15, 16}
impl_dot_array_for_slice!{17, 18, 19, 20, 21, 22, 23, 24}
impl_dot_array_for_slice!{25, 26, 27, 28, 29, 30, 31, 32}

/// In-place partitioning
pub trait Partition<T> {
    /// Partition `self` in place so that all elements for which the predicate holds are placed in
    /// the beginning. Return index of first element for which the predicate does not hold.
    fn partition<F: FnMut(&T) -> bool>(&mut self, predicate: F) -> usize;
}

impl<T> Partition<T> for [T]
{
    /// Partition a slice in place in O(n).
    fn partition<F: FnMut(&T) -> bool>(&mut self, mut predicate: F) -> usize {
        // Using unsafe pointer arithmetic is 50% faster than safe indexing
        unsafe {
            let start = &mut self[0] as *mut T;
            let end = start.offset(self.len() as isize);
            let mut target = &mut self[0] as *mut T;
            let mut current = target;
            let mut count = 0;
            while current < end {
                if predicate(&*current) {
                    mem::swap(&mut *target, &mut *current);
                    target = target.offset(1);
                    count += 1;
                }
                current = current.offset(1);
            }
            count
        }
    }
}

pub fn resample<T: Clone, R: Rng>(x: &[T], n: usize, rng: &mut R) -> Vec<T> {
    rng.sample_iter(&Uniform::new(0, x.len()))
        .take(n)
        .map(|i| x[i].clone())
        .collect()
}


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

    #[test]
    fn dot_vec() {
        let a = vec!(1.0, 2.0, 3.0);
        let b = vec!(2.0, 1.0, 1.0);
        assert_eq!(a.dot(&b), 7.0);

        let a = vec!(1, 2, 1);
        let b = vec!(2, 0, -3);
        assert_eq!(a.dot(&b), -1);

        let a = vec!(1, 2, -1);
        let b = vec!(2, 1, 4);
        assert_eq!(a.dot(&b), 0);
    }

    #[test]
    fn dot_array() {
        let a = [1.0, 2.0, 3.0];
        let b = [2.0, 1.0, 1.0];
        assert_eq!(a.dot(&b), 7.0);

        let a = [1, 2, 1];
        let b = [2, 0, -3];
        assert_eq!(a.dot(&b), -1);

        let a = [1, 2, -1];
        let b = [2, 1, 4];
        assert_eq!(a.dot(&b), 0);
    }

    #[test]
    fn dot_slice() {
        let a: &[_] = &[1.0, 2.0, 3.0];
        let b: &[_] = &[2.0, 1.0, 1.0];
        assert_eq!(a.dot(b), 7.0);

        let a: &[_] = &[1, 2, 1];
        let b: &[_] = &[2, 0, -3];
        assert_eq!(a.dot(b), -1);

        let a: &[_] = &[1, 2, -1];
        let b: &[_] = &[2, 1, 4];
        assert_eq!(a.dot(b), 0);
    }

    #[test]
    fn partition() {
        let mut x = [9, 8, 7, 6, 5, 4, 3, 2, 1, 0];
        let i = x.partition(|&xi| xi < 5);
        assert_eq!(i, 5);
        assert!(x[..i].iter().all(|&xi| xi < 5));
        assert!(x[i..].iter().all(|&xi| xi >= 5));

        // correctly partitioned sequence stays unchanged
        let mut x = vec![0, 1, 2, 3, 4, 5, 6, 7, 8, 9];
        let i = x.partition(|&xi| xi <= 3);
        assert_eq!(i, 4);
        assert_eq!(x, [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]);
    }

    #[test]
    fn bootstrap() {
        let x: Vec<_> = (1..=3).collect();
        let y = resample(&x, 1_000_000, &mut thread_rng());

        let (mut a, mut b, mut c) = (0.0, 0.0, 0.0);

        for z in &y {
            match z {
                1 => a += 1.0,
                2 => b += 1.0,
                3 => c += 1.0,
                _ => panic!("Value {} was not in original array", z)
            }
        }

        let expected = y.len() as f64 / 3.0;

        let chi_square = (a - expected) * (a - expected) / expected
            + (b - expected) * (b - expected) / expected
            + (c - expected) * (c - expected) / expected;

        // check critical value for chi-square distribution with 2 degrees
        // of freedom at the 99% significance level. (Note that this means
        // the test is expected to fail in 1% of runs)
        // https://www.itl.nist.gov/div898/handbook/eda/section3/eda3674.htm
        assert!(chi_square < 9.21);
    }
}