rill-ml 0.8.0

Lightweight, serializable online machine learning for Rust applications and streaming data.
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

//! Integration tests: regression model learning behavior.
//!
//! Verifies that `LinearRegression` and baseline regressors converge to
//! correct solutions on known data-generating processes.

use rand::SeedableRng;
use rill_ml::OnlineRegressor;
use rill_ml::loss::RegressionLoss;
use rill_ml::models::{
    BaselineConfig, ExponentiallyWeightedMeanRegressor, LastValueRegressor, LinearRegression,
    LinearRegressionConfig, MeanRegressor,
};
use rill_ml::optim::{Optimizer, SgdConfig};

#[test]
fn linear_regression_converges_to_true_weights() {
    let d = 3;
    let true_weights = [2.0, -1.0, 0.5];
    let true_intercept = 3.0;
    let mut model = LinearRegression::new(
        d,
        LinearRegressionConfig {
            optimizer: Optimizer::sgd(
                d,
                SgdConfig {
                    learning_rate: 0.05,
                    l2: 0.0,
                },
            )
            .unwrap(),
            loss: RegressionLoss::default(),
        },
    )
    .unwrap();

    let mut rng = rand_chacha::ChaCha8Rng::seed_from_u64(42);
    for _ in 0..2000 {
        let x: Vec<f64> = (0..d)
            .map(|_| rand::Rng::gen_range(&mut rng, -2.0..2.0))
            .collect();
        let y = true_weights
            .iter()
            .zip(&x)
            .map(|(w, xi)| w * xi)
            .sum::<f64>()
            + true_intercept;
        model.learn(&x, y).unwrap();
    }

    let test_x = vec![1.5, -0.5, 2.0];
    let expected_y: f64 = true_weights
        .iter()
        .zip(&test_x)
        .map(|(w, xi)| w * xi)
        .sum::<f64>()
        + true_intercept;
    let predicted = model.predict(&test_x).unwrap();
    assert!(
        (predicted - expected_y).abs() < 0.5,
        "predicted = {predicted}, expected = {expected_y}"
    );
}

#[test]
fn mean_regressor_converges_to_mean() {
    let mut model = MeanRegressor::new(BaselineConfig::default()).unwrap();
    let data = [10.0, 20.0, 30.0, 40.0, 50.0];
    for &y in &data {
        model.learn(&[], y).unwrap();
    }
    let predicted = model.predict(&[]).unwrap();
    assert!((predicted - 30.0).abs() < 1e-9);
}

#[test]
fn ew_mean_regressor_weights_recent() {
    let mut model =
        ExponentiallyWeightedMeanRegressor::new(0.5, BaselineConfig::default()).unwrap();
    model.learn(&[], 10.0).unwrap();
    model.learn(&[], 20.0).unwrap();
    model.learn(&[], 30.0).unwrap();
    let predicted = model.predict(&[]).unwrap();
    // EWMean: 10 -> 15 -> 22.5
    assert!((predicted - 22.5).abs() < 1e-9);
}

#[test]
fn last_value_regressor_tracks_last() {
    let mut model = LastValueRegressor::new(BaselineConfig::default()).unwrap();
    model.learn(&[], 7.0).unwrap();
    model.learn(&[], 42.0).unwrap();
    assert_eq!(model.predict(&[]).unwrap(), 42.0);
}

#[test]
fn linear_regression_with_adagrad_converges() {
    use rill_ml::optim::AdaGradConfig;
    let d = 2;
    let mut model = LinearRegression::new(
        d,
        LinearRegressionConfig {
            optimizer: Optimizer::adagrad(
                d,
                AdaGradConfig {
                    learning_rate: 1.0,
                    l2: 0.0,
                    epsilon: 1e-8,
                },
            )
            .unwrap(),
            loss: RegressionLoss::default(),
        },
    )
    .unwrap();

    let mut rng = rand_chacha::ChaCha8Rng::seed_from_u64(7);
    for _ in 0..1000 {
        let x1 = rand::Rng::gen_range(&mut rng, -1.0..1.0);
        let x2 = rand::Rng::gen_range(&mut rng, -1.0..1.0);
        let y = 1.0 * x1 + (-2.0) * x2;
        model.learn(&[x1, x2], y).unwrap();
    }

    let pred = model.predict(&[1.0, -1.0]).unwrap();
    // y = 1*1 + (-2)*(-1) = 3
    assert!((pred - 3.0).abs() < 0.5, "pred = {pred}");
}

#[test]
fn predict_is_side_effect_free() {
    let d = 2;
    let model = LinearRegression::new(
        d,
        LinearRegressionConfig {
            optimizer: Optimizer::sgd(d, SgdConfig::default()).unwrap(),
            loss: RegressionLoss::default(),
        },
    )
    .unwrap();
    let samples_before = model.samples_seen();
    let _ = model.predict(&[1.0, 2.0]).unwrap();
    let _ = model.predict(&[3.0, 4.0]).unwrap();
    assert_eq!(model.samples_seen(), samples_before);
}

#[test]
fn reset_clears_model_state() {
    let d = 1;
    let mut model = LinearRegression::new(
        d,
        LinearRegressionConfig {
            optimizer: Optimizer::sgd(d, SgdConfig::default()).unwrap(),
            loss: RegressionLoss::default(),
        },
    )
    .unwrap();
    model.learn(&[1.0], 5.0).unwrap();
    model.learn(&[2.0], 10.0).unwrap();
    assert_eq!(model.samples_seen(), 2);
    model.reset();
    assert_eq!(model.samples_seen(), 0);
    assert_eq!(model.predict(&[1.0]).unwrap(), 0.0);
}