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

use super::core::{RandomForest, RandomForestOptions};
use crate::criterion::ClassificationCriterion;
use crate::functions;
use crate::table::Table;
use std::io::{Read, Write};
use std::num::NonZeroUsize;

/// Random forest options.
#[derive(Debug, Clone, Default)]
pub struct RandomForestClassifierOptions {
    inner: RandomForestOptions,
}

impl RandomForestClassifierOptions {
    /// Makes a `RandomForestClassifierOptions` instance with the default settings.
    pub fn new() -> Self {
        Self::default()
    }

    /// Sets the random generator seed.
    ///
    /// The default value is random.
    pub fn seed(&mut self, seed: u64) -> &mut Self {
        self.inner.seed(seed);
        self
    }

    /// Sets the number of decision trees.
    ///
    /// The default value is `100`.
    pub fn trees(&mut self, trees: NonZeroUsize) -> &mut Self {
        self.inner.trees(trees);
        self
    }

    /// Sets the number of maximum candidate features used to determine each decision tree node.
    ///
    /// The default value is `sqrt(the number of features)`.
    pub fn max_features(&mut self, max: NonZeroUsize) -> &mut Self {
        self.inner.max_features(max);
        self
    }

    /// Sets the maximum number of samples used to train each decision tree.
    ///
    /// The default value is the number of rows in the target table.
    pub fn max_samples(&mut self, max: NonZeroUsize) -> &mut Self {
        self.inner.max_samples(max);
        self
    }

    /// Enables parallel executions of `RandomForestClassifier::fit`.
    ///
    /// This library use `rayon` for parallel execution.
    /// Please see [the rayon document](https://docs.rs/rayon) if you want to configure the behavior
    /// (e.g., the number of worker threads).
    pub fn parallel(&mut self) -> &mut Self {
        self.inner.parallel();
        self
    }

    /// Builds a classifier model fitting the given table.
    pub fn fit<T: ClassificationCriterion>(
        &self,
        criterion: T,
        table: Table,
    ) -> RandomForestClassifier {
        RandomForestClassifier {
            inner: self.inner.fit(criterion, false, table),
        }
    }
}

/// Random forest classifier.
#[derive(Debug)]
pub struct RandomForestClassifier {
    inner: RandomForest,
}

impl RandomForestClassifier {
    /// Builds a classifier model fitting the given table with the default settings.
    pub fn fit<T: ClassificationCriterion>(criterion: T, table: Table) -> Self {
        RandomForestClassifierOptions::default().fit(criterion, table)
    }

    /// Predicts the target value for the given features.
    pub fn predict(&self, features: &[f64]) -> f64 {
        functions::most_frequent(self.inner.predict(features))
    }

    /// Writes this classifier to the given byte stream.
    pub fn serialize<W: Write>(&self, writer: W) -> std::io::Result<()> {
        self.inner.serialize(writer)
    }

    /// Reads a classifier from the given byte stream.
    pub fn deserialize<R: Read>(reader: R) -> std::io::Result<Self> {
        let inner = RandomForest::deserialize(reader)?;
        Ok(Self { inner })
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::criterion::Gini;
    use crate::table::{ColumnType, TableBuilder};

    #[test]
    fn classification_works() -> Result<(), anyhow::Error> {
        let features = [
            &[0.0, 1.0, 0.0][..],
            &[1.0, 1.0, 1.0][..],
            &[0.0, 0.0, 1.0][..],
            &[1.0, 0.0, 0.0][..],
            &[1.0, 0.0, 0.0][..],
            &[0.0, 1.0, 0.0][..],
            &[1.0, 0.0, 1.0][..],
        ];

        let target = [1.0, 0.0, 1.0, 0.0, 1.0, 1.0, 0.0];
        let train_len = target.len() - 1;

        let mut table_builder = TableBuilder::new();
        table_builder.set_feature_column_types(&[
            ColumnType::Categorical,
            ColumnType::Categorical,
            ColumnType::Categorical,
        ])?;

        for (xs, y) in features.iter().zip(target.iter()).take(train_len) {
            table_builder.add_row(xs, *y)?;
        }
        let table = table_builder.build()?;

        let classifier = RandomForestClassifierOptions::new()
            .seed(0)
            .fit(Gini, table.clone());
        assert_eq!(classifier.predict(&features[train_len]), 0.0);

        let classifier_parallel = RandomForestClassifierOptions::new()
            .seed(0)
            .parallel()
            .fit(Gini, table);
        assert_eq!(
            classifier.predict(&features[train_len]),
            classifier_parallel.predict(&features[train_len])
        );

        let mut bytes = Vec::new();
        classifier.serialize(&mut bytes)?;
        let classifier_deserialized = RandomForestClassifier::deserialize(&mut &bytes[..])?;
        assert_eq!(
            classifier.predict(&features[train_len]),
            classifier_deserialized.predict(&features[train_len])
        );

        Ok(())
    }
}