work/smartcore
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Port ensemble. Add Display to naive_bayes (#208)

Browse Source
This commit is contained in:
Lorenzo
2022-10-31 17:35:33 +00:00
committed by morenol
parent d91f4f7ce4
commit a16927aa16
10 changed files with 330 additions and 242 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/coverage.yml
+1 -1
View File
@@ -39,6 +39,6 @@ jobs:
command: tarpaulin
args: --out Lcov --all-features -- --test-threads 1
- name: Upload to codecov.io
uses: codecov/codecov-action@v1
uses: codecov/codecov-action@v2
with:
fail_ci_if_error: true
src/ensemble/random_forest_classifier.rs
+108 -78
View File
@@ -54,7 +54,7 @@ use std::fmt::Debug;
use serde::{Deserialize, Serialize};
use crate::api::{Predictor, SupervisedEstimator};
use crate::error::Failed;
use crate::error::{Failed, FailedError};
use crate::linalg::basic::arrays::{Array1, Array2};
use crate::numbers::basenum::Number;
use crate::numbers::floatnum::FloatNumber;
@@ -104,9 +104,10 @@ pub struct RandomForestClassifier<
X: Array2<TX>,
Y: Array1<TY>,
> {
parameters: RandomForestClassifierParameters,
trees: Vec<DecisionTreeClassifier<TX, TY, X, Y>>,
classes: Vec<TY>,
parameters: Option<RandomForestClassifierParameters>,
trees: Option<Vec<DecisionTreeClassifier<TX, TY, X, Y>>>,
classes: Option<Vec<TY>>,
samples: Option<Vec<Vec<bool>>>,
}
impl RandomForestClassifierParameters {
@@ -154,11 +155,13 @@ impl RandomForestClassifierParameters {
}
}
impl<TX: Number + FloatNumber + PartialOrd, TY: Number + Ord, X: Array2<TX>, Y: Array1<TY>> PartialEq
for RandomForestClassifier<TX, TY, X, Y>
impl<TX: Number + FloatNumber + PartialOrd, TY: Number + Ord, X: Array2<TX>, Y: Array1<TY>>
PartialEq for RandomForestClassifier<TX, TY, X, Y>
{
fn eq(&self, other: &Self) -> bool {
if self.classes.len() != other.classes.len() || self.trees.len() != other.trees.len() {
if self.classes.as_ref().unwrap().len() != other.classes.as_ref().unwrap().len()
|| self.trees.as_ref().unwrap().len() != other.trees.as_ref().unwrap().len()
{
false
} else {
self.classes
@@ -189,17 +192,25 @@ impl Default for RandomForestClassifierParameters {
}
}
impl<TX: FloatNumber + PartialOrd, TY: Number + Ord, X: Array2<TX>, Y: Array1<TY>>
impl<TX: Number + FloatNumber + PartialOrd, TY: Number + Ord, X: Array2<TX>, Y: Array1<TY>>
SupervisedEstimator<X, Y, RandomForestClassifierParameters>
for RandomForestClassifier<TX, TY, X, Y>
{
fn new() -> Self {
Self {
parameters: Option::None,
trees: Option::None,
classes: Option::None,
samples: Option::None,
}
}
fn fit(x: &X, y: &Y, parameters: RandomForestClassifierParameters) -> Result<Self, Failed> {
RandomForestClassifier::fit(x, y, parameters)
}
}
impl<TX: Number + FloatNumber + PartialOrd, TY: Number + Ord, X: Array2<TX>, Y: Array1<TY>> Predictor<X, Y>
for RandomForestClassifier<TX, TY, X, Y>
impl<TX: Number + FloatNumber + PartialOrd, TY: Number + Ord, X: Array2<TX>, Y: Array1<TY>>
Predictor<X, Y> for RandomForestClassifier<TX, TY, X, Y>
{
fn predict(&self, x: &X) -> Result<Y, Failed> {
self.predict(x)
@@ -462,10 +473,22 @@ impl<TX: FloatNumber + PartialOrd, TY: Number + Ord, X: Array2<TX>, Y: Array1<TY
let mut rng = get_rng_impl(Some(parameters.seed));
let classes = y.unique();
let k = classes.len();
// TODO: use with_capacity here
let mut trees: Vec<DecisionTreeClassifier<TX, TY, X, Y>> = Vec::new();
let mut maybe_all_samples: Option<Vec<Vec<bool>>> = Option::None;
if parameters.keep_samples {
// TODO: use with_capacity here
maybe_all_samples = Some(Vec::new());
}
for _ in 0..parameters.n_trees {
let samples = RandomForestClassifier::<TX, TY, X, Y>::sample_with_replacement(&yi, k, &mut rng);
let samples: Vec<usize> =
RandomForestClassifier::<TX, TY, X, Y>::sample_with_replacement(&yi, k, &mut rng);
if let Some(ref mut all_samples) = maybe_all_samples {
all_samples.push(samples.iter().map(|x| *x != 0).collect())
}
let params = DecisionTreeClassifierParameters {
criterion: parameters.criterion.clone(),
max_depth: parameters.max_depth,
@@ -478,9 +501,10 @@ impl<TX: FloatNumber + PartialOrd, TY: Number + Ord, X: Array2<TX>, Y: Array1<TY
}
Ok(RandomForestClassifier {
parameters: parameters,
trees,
classes,
parameters: Some(parameters),
trees: Some(trees),
classes: Some(classes),
samples: maybe_all_samples,
})
}
@@ -492,16 +516,19 @@ impl<TX: FloatNumber + PartialOrd, TY: Number + Ord, X: Array2<TX>, Y: Array1<TY
let (n, _) = x.shape();
for i in 0..n {
result.set(i, self.classes[self.predict_for_row(x, i)]);
result.set(
i,
self.classes.as_ref().unwrap()[self.predict_for_row(x, i)],
);
}
Ok(result)
}
fn predict_for_row(&self, x: &X, row: usize) -> usize {
let mut result = vec![0; self.classes.len()];
let mut result = vec![0; self.classes.as_ref().unwrap().len()];
for tree in self.trees.iter() {
for tree in self.trees.as_ref().unwrap().iter() {
result[tree.predict_for_row(x, row)] += 1;
}
@@ -511,38 +538,43 @@ impl<TX: FloatNumber + PartialOrd, TY: Number + Ord, X: Array2<TX>, Y: Array1<TY
/// Predict OOB classes for `x`. `x` is expected to be equal to the dataset used in training.
pub fn predict_oob(&self, x: &X) -> Result<Y, Failed> {
let (n, _) = x.shape();
/* TODO: fix this:
if self.samples.is_none() {
Err(Failed::because(
FailedError::PredictFailed,
"Need samples=true for OOB predictions.",
))
} else if self.samples.as_ref().unwrap()[0].len() != n {
Err(Failed::because(
FailedError::PredictFailed,
"Prediction matrix must match matrix used in training for OOB predictions.",
))
} else {
*/
let mut result = Y::zeros(n);
if self.samples.is_none() {
Err(Failed::because(
FailedError::PredictFailed,
"Need samples=true for OOB predictions.",
))
} else if self.samples.as_ref().unwrap()[0].len() != n {
Err(Failed::because(
FailedError::PredictFailed,
"Prediction matrix must match matrix used in training for OOB predictions.",
))
} else {
let mut result = Y::zeros(n);
for i in 0..n {
result.set(i, self.classes[self.predict_for_row_oob(x, i)]);
for i in 0..n {
result.set(
i,
self.classes.as_ref().unwrap()[self.predict_for_row_oob(x, i)],
);
}
Ok(result)
}
Ok(result)
//}
}
fn predict_for_row_oob(&self, x: &X, row: usize) -> usize {
let mut result = vec![0; self.classes.len()];
let mut result = vec![0; self.classes.as_ref().unwrap().len()];
// TODO: FIX THIS
//for (tree, samples) in self.trees.iter().zip(self.samples.as_ref().unwrap()) {
// if !samples[row] {
// result[tree.predict_for_row(x, row)] += 1;
// }
// }
for (tree, samples) in self
.trees
.as_ref()
.unwrap()
.iter()
.zip(self.samples.as_ref().unwrap())
{
if !samples[row] {
result[tree.predict_for_row(x, row)] += 1;
}
}
which_max(&result)
}
@@ -671,9 +703,7 @@ mod tests {
&[6.6, 2.9, 4.6, 1.3],
&[5.2, 2.7, 3.9, 1.4],
]);
let y = vec![
0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
];
let y = vec![0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1];
let classifier = RandomForestClassifier::fit(
&x,
@@ -697,39 +727,39 @@ mod tests {
);
}
// #[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
// #[test]
// #[cfg(feature = "serde")]
// fn serde() {
// let x = DenseMatrix::from_2d_array(&[
// &[5.1, 3.5, 1.4, 0.2],
// &[4.9, 3.0, 1.4, 0.2],
// &[4.7, 3.2, 1.3, 0.2],
// &[4.6, 3.1, 1.5, 0.2],
// &[5.0, 3.6, 1.4, 0.2],
// &[5.4, 3.9, 1.7, 0.4],
// &[4.6, 3.4, 1.4, 0.3],
// &[5.0, 3.4, 1.5, 0.2],
// &[4.4, 2.9, 1.4, 0.2],
// &[4.9, 3.1, 1.5, 0.1],
// &[7.0, 3.2, 4.7, 1.4],
// &[6.4, 3.2, 4.5, 1.5],
// &[6.9, 3.1, 4.9, 1.5],
// &[5.5, 2.3, 4.0, 1.3],
// &[6.5, 2.8, 4.6, 1.5],
// &[5.7, 2.8, 4.5, 1.3],
// &[6.3, 3.3, 4.7, 1.6],
// &[4.9, 2.4, 3.3, 1.0],
// &[6.6, 2.9, 4.6, 1.3],
// &[5.2, 2.7, 3.9, 1.4],
// ]);
// let y = vec![0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1];
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
#[cfg(feature = "serde")]
fn serde() {
let x = DenseMatrix::from_2d_array(&[
&[5.1, 3.5, 1.4, 0.2],
&[4.9, 3.0, 1.4, 0.2],
&[4.7, 3.2, 1.3, 0.2],
&[4.6, 3.1, 1.5, 0.2],
&[5.0, 3.6, 1.4, 0.2],
&[5.4, 3.9, 1.7, 0.4],
&[4.6, 3.4, 1.4, 0.3],
&[5.0, 3.4, 1.5, 0.2],
&[4.4, 2.9, 1.4, 0.2],
&[4.9, 3.1, 1.5, 0.1],
&[7.0, 3.2, 4.7, 1.4],
&[6.4, 3.2, 4.5, 1.5],
&[6.9, 3.1, 4.9, 1.5],
&[5.5, 2.3, 4.0, 1.3],
&[6.5, 2.8, 4.6, 1.5],
&[5.7, 2.8, 4.5, 1.3],
&[6.3, 3.3, 4.7, 1.6],
&[4.9, 2.4, 3.3, 1.0],
&[6.6, 2.9, 4.6, 1.3],
&[5.2, 2.7, 3.9, 1.4],
]);
let y = vec![0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1];
// let forest = RandomForestClassifier::fit(&x, &y, Default::default()).unwrap();
let forest = RandomForestClassifier::fit(&x, &y, Default::default()).unwrap();
// let deserialized_forest: RandomForestClassifier<f64, i64, DenseMatrix<f64>, Vec<i64>> =
// bincode::deserialize(&bincode::serialize(&forest).unwrap()).unwrap();
let deserialized_forest: RandomForestClassifier<f64, i64, DenseMatrix<f64>, Vec<i64>> =
bincode::deserialize(&bincode::serialize(&forest).unwrap()).unwrap();
// assert_eq!(forest, deserialized_forest);
// }
assert_eq!(forest, deserialized_forest);
}
}
src/ensemble/random_forest_regressor.rs
+83 -75
View File
@@ -51,7 +51,7 @@ use std::fmt::Debug;
use serde::{Deserialize, Serialize};
use crate::api::{Predictor, SupervisedEstimator};
use crate::error::Failed;
use crate::error::{Failed, FailedError};
use crate::linalg::basic::arrays::{Array1, Array2};
use crate::numbers::basenum::Number;
use crate::numbers::floatnum::FloatNumber;
@@ -92,11 +92,15 @@ pub struct RandomForestRegressorParameters {
/// Random Forest Regressor
#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
#[derive(Debug)]
pub struct RandomForestRegressor<TX: Number + FloatNumber + PartialOrd, TY: Number, X: Array2<TX>, Y: Array1<TY>>
{
parameters: RandomForestRegressorParameters,
trees: Vec<DecisionTreeRegressor<TX, TY, X, Y>>,
samples: Option<Vec<Vec<usize>>>
pub struct RandomForestRegressor<
TX: Number + FloatNumber + PartialOrd,
TY: Number,
X: Array2<TX>,
Y: Array1<TY>,
> {
parameters: Option<RandomForestRegressorParameters>,
trees: Option<Vec<DecisionTreeRegressor<TX, TY, X, Y>>>,
samples: Option<Vec<Vec<bool>>>,
}
impl RandomForestRegressorParameters {
@@ -156,7 +160,7 @@ impl<TX: Number + FloatNumber + PartialOrd, TY: Number, X: Array2<TX>, Y: Array1
for RandomForestRegressor<TX, TY, X, Y>
{
fn eq(&self, other: &Self) -> bool {
if self.trees.len() != other.trees.len() {
if self.trees.as_ref().unwrap().len() != other.trees.as_ref().unwrap().len() {
false
} else {
self.trees
@@ -171,13 +175,21 @@ impl<TX: Number + FloatNumber + PartialOrd, TY: Number, X: Array2<TX>, Y: Array1
SupervisedEstimator<X, Y, RandomForestRegressorParameters>
for RandomForestRegressor<TX, TY, X, Y>
{
fn new() -> Self {
Self {
parameters: Option::None,
trees: Option::None,
samples: Option::None,
}
}
fn fit(x: &X, y: &Y, parameters: RandomForestRegressorParameters) -> Result<Self, Failed> {
RandomForestRegressor::fit(x, y, parameters)
}
}
impl<TX: Number + FloatNumber + PartialOrd, TY: Number, X: Array2<TX>, Y: Array1<TY>> Predictor<X, Y>
for RandomForestRegressor<TX, TY, X, Y>
impl<TX: Number + FloatNumber + PartialOrd, TY: Number, X: Array2<TX>, Y: Array1<TY>>
Predictor<X, Y> for RandomForestRegressor<TX, TY, X, Y>
{
fn predict(&self, x: &X) -> Result<Y, Failed> {
self.predict(x)
@@ -396,17 +408,19 @@ impl<TX: Number + FloatNumber + PartialOrd, TY: Number, X: Array2<TX>, Y: Array1
let mut rng = get_rng_impl(Some(parameters.seed));
let mut trees: Vec<DecisionTreeRegressor<TX, TY, X, Y>> = Vec::new();
let mut maybe_all_samples: Vec<Vec<usize>> = Vec::new();
let mut maybe_all_samples: Option<Vec<Vec<bool>>> = Option::None;
if parameters.keep_samples {
// TODO: use with_capacity here
maybe_all_samples = Some(Vec::new());
}
for _ in 0..parameters.n_trees {
let samples = RandomForestRegressor::<TX, TY, X, Y>::sample_with_replacement(
n_rows,
&mut rng,
);
let samples: Vec<usize> =
RandomForestRegressor::<TX, TY, X, Y>::sample_with_replacement(n_rows, &mut rng);
// keep samples is flag is on
if parameters.keep_samples {
maybe_all_samples.push(samples);
if let Some(ref mut all_samples) = maybe_all_samples {
all_samples.push(samples.iter().map(|x| *x != 0).collect())
}
let params = DecisionTreeRegressorParameters {
@@ -419,17 +433,10 @@ impl<TX: Number + FloatNumber + PartialOrd, TY: Number, X: Array2<TX>, Y: Array1
trees.push(tree);
}
let samples;
if maybe_all_samples.len() == 0 {
samples = Option::None;
} else {
samples = Some(maybe_all_samples)
}
Ok(RandomForestRegressor {
parameters: parameters,
trees,
samples
parameters: Some(parameters),
trees: Some(trees),
samples: maybe_all_samples,
})
}
@@ -448,11 +455,11 @@ impl<TX: Number + FloatNumber + PartialOrd, TY: Number, X: Array2<TX>, Y: Array1
}
fn predict_for_row(&self, x: &X, row: usize) -> TY {
let n_trees = self.trees.len();
let n_trees = self.trees.as_ref().unwrap().len();
let mut result = TY::zero();
for tree in self.trees.iter() {
for tree in self.trees.as_ref().unwrap().iter() {
result += tree.predict_for_row(x, row);
}
@@ -462,7 +469,6 @@ impl<TX: Number + FloatNumber + PartialOrd, TY: Number, X: Array2<TX>, Y: Array1
/// Predict OOB classes for `x`. `x` is expected to be equal to the dataset used in training.
pub fn predict_oob(&self, x: &X) -> Result<Y, Failed> {
let (n, _) = x.shape();
/* TODO: FIX THIS
if self.samples.is_none() {
Err(Failed::because(
FailedError::PredictFailed,
@@ -474,28 +480,31 @@ impl<TX: Number + FloatNumber + PartialOrd, TY: Number, X: Array2<TX>, Y: Array1
"Prediction matrix must match matrix used in training for OOB predictions.",
))
} else {
let mut result = Y::zeros(n);
let mut result = Y::zeros(n);
for i in 0..n {
result.set(i, self.predict_for_row_oob(x, i));
for i in 0..n {
result.set(i, self.predict_for_row_oob(x, i));
}
Ok(result)
}
Ok(result)
}*/
let result = Y::zeros(n);
Ok(result)
}
//TODo: fix this
fn predict_for_row_oob(&self, x: &X, row: usize) -> TY {
let mut n_trees = 0;
let mut result = TY::zero();
for (tree, samples) in self.trees.iter().zip(self.samples.as_ref().unwrap()) {
if !samples[row] {
result += tree.predict_for_row(x, row);
n_trees += 1;
}
for (tree, samples) in self
.trees
.as_ref()
.unwrap()
.iter()
.zip(self.samples.as_ref().unwrap())
{
if !samples[row] {
result += tree.predict_for_row(x, row);
n_trees += 1;
}
}
// TODO: What to do if there are no oob trees?
@@ -636,39 +645,38 @@ mod tests {
assert!(mean_absolute_error(&y, &y_hat) < mean_absolute_error(&y, &y_hat_oob));
}
// TODO: missing deserialization for DenseMatrix
// #[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
// #[test]
// #[cfg(feature = "serde")]
// fn serde() {
// let x = DenseMatrix::from_2d_array(&[
// &[234.289, 235.6, 159., 107.608, 1947., 60.323],
// &[259.426, 232.5, 145.6, 108.632, 1948., 61.122],
// &[258.054, 368.2, 161.6, 109.773, 1949., 60.171],
// &[284.599, 335.1, 165., 110.929, 1950., 61.187],
// &[328.975, 209.9, 309.9, 112.075, 1951., 63.221],
// &[346.999, 193.2, 359.4, 113.27, 1952., 63.639],
// &[365.385, 187., 354.7, 115.094, 1953., 64.989],
// &[363.112, 357.8, 335., 116.219, 1954., 63.761],
// &[397.469, 290.4, 304.8, 117.388, 1955., 66.019],
// &[419.18, 282.2, 285.7, 118.734, 1956., 67.857],
// &[442.769, 293.6, 279.8, 120.445, 1957., 68.169],
// &[444.546, 468.1, 263.7, 121.95, 1958., 66.513],
// &[482.704, 381.3, 255.2, 123.366, 1959., 68.655],
// &[502.601, 393.1, 251.4, 125.368, 1960., 69.564],
// &[518.173, 480.6, 257.2, 127.852, 1961., 69.331],
// &[554.894, 400.7, 282.7, 130.081, 1962., 70.551],
// ]);
// let y = vec![
// 83.0, 88.5, 88.2, 89.5, 96.2, 98.1, 99.0, 100.0, 101.2, 104.6, 108.4, 110.8, 112.6,
// 114.2, 115.7, 116.9,
// ];
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
#[cfg(feature = "serde")]
fn serde() {
let x = DenseMatrix::from_2d_array(&[
&[234.289, 235.6, 159., 107.608, 1947., 60.323],
&[259.426, 232.5, 145.6, 108.632, 1948., 61.122],
&[258.054, 368.2, 161.6, 109.773, 1949., 60.171],
&[284.599, 335.1, 165., 110.929, 1950., 61.187],
&[328.975, 209.9, 309.9, 112.075, 1951., 63.221],
&[346.999, 193.2, 359.4, 113.27, 1952., 63.639],
&[365.385, 187., 354.7, 115.094, 1953., 64.989],
&[363.112, 357.8, 335., 116.219, 1954., 63.761],
&[397.469, 290.4, 304.8, 117.388, 1955., 66.019],
&[419.18, 282.2, 285.7, 118.734, 1956., 67.857],
&[442.769, 293.6, 279.8, 120.445, 1957., 68.169],
&[444.546, 468.1, 263.7, 121.95, 1958., 66.513],
&[482.704, 381.3, 255.2, 123.366, 1959., 68.655],
&[502.601, 393.1, 251.4, 125.368, 1960., 69.564],
&[518.173, 480.6, 257.2, 127.852, 1961., 69.331],
&[554.894, 400.7, 282.7, 130.081, 1962., 70.551],
]);
let y = vec![
83.0, 88.5, 88.2, 89.5, 96.2, 98.1, 99.0, 100.0, 101.2, 104.6, 108.4, 110.8, 112.6,
114.2, 115.7, 116.9,
];
// let forest = RandomForestRegressor::fit(&x, &y, Default::default()).unwrap();
let forest = RandomForestRegressor::fit(&x, &y, Default::default()).unwrap();
// let deserialized_forest: RandomForestRegressor<f64, f64, DenseMatrix<f64>, Vec<f64>> =
// bincode::deserialize(&bincode::serialize(&forest).unwrap()).unwrap();
let deserialized_forest: RandomForestRegressor<f64, f64, DenseMatrix<f64>, Vec<f64>> =
bincode::deserialize(&bincode::serialize(&forest).unwrap()).unwrap();
// assert_eq!(forest, deserialized_forest);
// }
assert_eq!(forest, deserialized_forest);
}
}
src/lib.rs
+1 -1
View File
@@ -80,7 +80,7 @@ pub mod dataset;
/// Matrix decomposition algorithms
pub mod decomposition;
/// Ensemble methods, including Random Forest classifier and regressor
// pub mod ensemble;
pub mod ensemble;
pub mod error;
/// Diverse collection of linear algebra abstractions and methods that power SmartCore algorithms
pub mod linalg;
src/linalg/basic/matrix.rs
+2 -2
View File
@@ -4,7 +4,7 @@ use std::ops::Range;
use std::slice::Iter;
use approx::{AbsDiffEq, RelativeEq};
use serde::Serialize;
use serde::{Deserialize, Serialize};
use crate::linalg::basic::arrays::{
Array, Array2, ArrayView1, ArrayView2, MutArray, MutArrayView2,
@@ -19,7 +19,7 @@ use crate::numbers::basenum::Number;
use crate::numbers::realnum::RealNumber;
/// Dense matrix
#[derive(Debug, Clone, Serialize)]
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct DenseMatrix<T> {
ncols: usize,
nrows: usize,
src/naive_bayes/bernoulli.rs
+30 -17
View File
@@ -33,6 +33,8 @@
//! ## References:
//!
//! * ["Introduction to Information Retrieval", Manning C. D., Raghavan P., Schutze H., 2009, Chapter 13 ](https://nlp.stanford.edu/IR-book/information-retrieval-book.html)
use std::fmt;
use num_traits::Unsigned;
use crate::api::{Predictor, SupervisedEstimator};
@@ -62,6 +64,18 @@ struct BernoulliNBDistribution<T: Number + Ord + Unsigned> {
n_features: usize,
}
impl<T: Number + Ord + Unsigned> fmt::Display for BernoulliNBDistribution<T> {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
writeln!(
f,
"BernoulliNBDistribution: n_features: {:?}",
self.n_features
)?;
writeln!(f, "class_labels: {:?}", self.class_labels)?;
Ok(())
}
}
impl<T: Number + Ord + Unsigned> PartialEq for BernoulliNBDistribution<T> {
fn eq(&self, other: &Self) -> bool {
if self.class_labels == other.class_labels
@@ -598,23 +612,22 @@ mod tests {
assert_eq!(y_hat, vec!(2, 2, 0, 0, 0, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0));
}
// TODO: implement serialization
// #[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
// #[test]
// #[cfg(feature = "serde")]
// fn serde() {
// let x = DenseMatrix::from_2d_array(&[
// &[1, 1, 0, 0, 0, 0],
// &[0, 1, 0, 0, 1, 0],
// &[0, 1, 0, 1, 0, 0],
// &[0, 1, 1, 0, 0, 1],
// ]);
// let y: Vec<u32> = vec![0, 0, 0, 1];
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
#[cfg(feature = "serde")]
fn serde() {
let x = DenseMatrix::from_2d_array(&[
&[1, 1, 0, 0, 0, 0],
&[0, 1, 0, 0, 1, 0],
&[0, 1, 0, 1, 0, 0],
&[0, 1, 1, 0, 0, 1],
]);
let y: Vec<u32> = vec![0, 0, 0, 1];
// let bnb = BernoulliNB::fit(&x, &y, Default::default()).unwrap();
// let deserialized_bnb: BernoulliNB<i32, u32, DenseMatrix<i32>, Vec<u32>> =
// serde_json::from_str(&serde_json::to_string(&bnb).unwrap()).unwrap();
let bnb = BernoulliNB::fit(&x, &y, Default::default()).unwrap();
let deserialized_bnb: BernoulliNB<i32, u32, DenseMatrix<i32>, Vec<u32>> =
serde_json::from_str(&serde_json::to_string(&bnb).unwrap()).unwrap();
// assert_eq!(bnb, deserialized_bnb);
// }
assert_eq!(bnb, deserialized_bnb);
}
}
src/naive_bayes/categorical.rs
+40 -27
View File
@@ -30,6 +30,8 @@
//! let nb = CategoricalNB::fit(&x, &y, Default::default()).unwrap();
//! let y_hat = nb.predict(&x).unwrap();
//! ```
use std::fmt;
use num_traits::Unsigned;
use crate::api::{Predictor, SupervisedEstimator};
@@ -61,6 +63,18 @@ struct CategoricalNBDistribution<T: Number + Unsigned> {
category_count: Vec<Vec<Vec<usize>>>,
}
impl<T: Number + Ord + Unsigned> fmt::Display for CategoricalNBDistribution<T> {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
writeln!(
f,
"CategoricalNBDistribution: n_features: {:?}",
self.n_features
)?;
writeln!(f, "class_labels: {:?}", self.class_labels)?;
Ok(())
}
}
impl<T: Number + Unsigned> PartialEq for CategoricalNBDistribution<T> {
fn eq(&self, other: &Self) -> bool {
if self.class_labels == other.class_labels
@@ -521,34 +535,33 @@ mod tests {
assert_eq!(y_hat, vec![0, 0, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1]);
}
// TODO: implement serialization
// #[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
// #[test]
// #[cfg(feature = "serde")]
// fn serde() {
// let x = DenseMatrix::from_2d_array(&[
// &[3, 4, 0, 1],
// &[3, 0, 0, 1],
// &[4, 4, 1, 2],
// &[4, 2, 4, 3],
// &[4, 2, 4, 2],
// &[4, 1, 1, 0],
// &[1, 1, 1, 1],
// &[0, 4, 1, 0],
// &[0, 3, 2, 1],
// &[0, 3, 1, 1],
// &[3, 4, 0, 1],
// &[3, 4, 2, 4],
// &[0, 3, 1, 2],
// &[0, 4, 1, 2],
// ]);
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
#[cfg(feature = "serde")]
fn serde() {
let x = DenseMatrix::from_2d_array(&[
&[3, 4, 0, 1],
&[3, 0, 0, 1],
&[4, 4, 1, 2],
&[4, 2, 4, 3],
&[4, 2, 4, 2],
&[4, 1, 1, 0],
&[1, 1, 1, 1],
&[0, 4, 1, 0],
&[0, 3, 2, 1],
&[0, 3, 1, 1],
&[3, 4, 0, 1],
&[3, 4, 2, 4],
&[0, 3, 1, 2],
&[0, 4, 1, 2],
]);
// let y: Vec<u32> = vec![0, 0, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 0];
// let cnb = CategoricalNB::fit(&x, &y, Default::default()).unwrap();
let y: Vec<u32> = vec![0, 0, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 0];
let cnb = CategoricalNB::fit(&x, &y, Default::default()).unwrap();
// let deserialized_cnb: CategoricalNB<u32, DenseMatrix<u32>, Vec<u32>> =
// serde_json::from_str(&serde_json::to_string(&cnb).unwrap()).unwrap();
let deserialized_cnb: CategoricalNB<u32, DenseMatrix<u32>, Vec<u32>> =
serde_json::from_str(&serde_json::to_string(&cnb).unwrap()).unwrap();
// assert_eq!(cnb, deserialized_cnb);
// }
assert_eq!(cnb, deserialized_cnb);
}
}
src/naive_bayes/gaussian.rs
+32 -19
View File
@@ -22,6 +22,8 @@
//! let nb = GaussianNB::fit(&x, &y, Default::default()).unwrap();
//! let y_hat = nb.predict(&x).unwrap();
//! ```
use std::fmt;
use num_traits::Unsigned;
use crate::api::{Predictor, SupervisedEstimator};
@@ -49,6 +51,18 @@ struct GaussianNBDistribution<T: Number> {
theta: Vec<Vec<f64>>,
}
impl<T: Number + Ord + Unsigned> fmt::Display for GaussianNBDistribution<T> {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
writeln!(
f,
"GaussianNBDistribution: class_count: {:?}",
self.class_count
)?;
writeln!(f, "class_labels: {:?}", self.class_labels)?;
Ok(())
}
}
impl<X: Number + RealNumber, Y: Number + Ord + Unsigned> NBDistribution<X, Y>
for GaussianNBDistribution<Y>
{
@@ -415,25 +429,24 @@ mod tests {
assert_eq!(gnb.class_priors(), &priors);
}
// TODO: implement serialization
// #[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
// #[test]
// #[cfg(feature = "serde")]
// fn serde() {
// let x = DenseMatrix::<f64>::from_2d_array(&[
// &[-1., -1.],
// &[-2., -1.],
// &[-3., -2.],
// &[1., 1.],
// &[2., 1.],
// &[3., 2.],
// ]);
// let y: Vec<u32> = vec![1, 1, 1, 2, 2, 2];
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
#[cfg(feature = "serde")]
fn serde() {
let x = DenseMatrix::<f64>::from_2d_array(&[
&[-1., -1.],
&[-2., -1.],
&[-3., -2.],
&[1., 1.],
&[2., 1.],
&[3., 2.],
]);
let y: Vec<u32> = vec![1, 1, 1, 2, 2, 2];
// let gnb = GaussianNB::fit(&x, &y, Default::default()).unwrap();
// let deserialized_gnb: GaussianNB<f64, u32, DenseMatrix<f64>, Vec<u32>> =
// serde_json::from_str(&serde_json::to_string(&gnb).unwrap()).unwrap();
let gnb = GaussianNB::fit(&x, &y, Default::default()).unwrap();
let deserialized_gnb: GaussianNB<f64, u32, DenseMatrix<f64>, Vec<u32>> =
serde_json::from_str(&serde_json::to_string(&gnb).unwrap()).unwrap();
// assert_eq!(gnb, deserialized_gnb);
// }
assert_eq!(gnb, deserialized_gnb);
}
}
src/naive_bayes/multinomial.rs
+30 -17
View File
@@ -33,6 +33,8 @@
//! ## References:
//!
//! * ["Introduction to Information Retrieval", Manning C. D., Raghavan P., Schutze H., 2009, Chapter 13 ](https://nlp.stanford.edu/IR-book/information-retrieval-book.html)
use std::fmt;
use num_traits::Unsigned;
use crate::api::{Predictor, SupervisedEstimator};
@@ -62,6 +64,18 @@ struct MultinomialNBDistribution<T: Number> {
n_features: usize,
}
impl<T: Number + Ord + Unsigned> fmt::Display for MultinomialNBDistribution<T> {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
writeln!(
f,
"MultinomialNBDistribution: n_features: {:?}",
self.n_features
)?;
writeln!(f, "class_labels: {:?}", self.class_labels)?;
Ok(())
}
}
impl<X: Number + Unsigned, Y: Number + Ord + Unsigned> NBDistribution<X, Y>
for MultinomialNBDistribution<Y>
{
@@ -510,23 +524,22 @@ mod tests {
assert_eq!(y_hat, vec!(2, 2, 0, 0, 0, 2, 2, 1, 0, 1, 0, 2, 0, 0, 2));
}
// TODO: implement serialization
// #[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
// #[test]
// #[cfg(feature = "serde")]
// fn serde() {
// let x = DenseMatrix::from_2d_array(&[
// &[1, 1, 0, 0, 0, 0],
// &[0, 1, 0, 0, 1, 0],
// &[0, 1, 0, 1, 0, 0],
// &[0, 1, 1, 0, 0, 1],
// ]);
// let y = vec![0, 0, 0, 1];
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
#[cfg(feature = "serde")]
fn serde() {
let x = DenseMatrix::from_2d_array(&[
&[1, 1, 0, 0, 0, 0],
&[0, 1, 0, 0, 1, 0],
&[0, 1, 0, 1, 0, 0],
&[0, 1, 1, 0, 0, 1],
]);
let y = vec![0, 0, 0, 1];
// let mnb = MultinomialNB::fit(&x, &y, Default::default()).unwrap();
// let deserialized_mnb: MultinomialNB<u32, u32, DenseMatrix<u32>, Vec<u32>> =
// serde_json::from_str(&serde_json::to_string(&mnb).unwrap()).unwrap();
let mnb = MultinomialNB::fit(&x, &y, Default::default()).unwrap();
let deserialized_mnb: MultinomialNB<u32, u32, DenseMatrix<u32>, Vec<u32>> =
serde_json::from_str(&serde_json::to_string(&mnb).unwrap()).unwrap();
// assert_eq!(mnb, deserialized_mnb);
// }
assert_eq!(mnb, deserialized_mnb);
}
}
src/svm/svc.rs
+2 -4
View File
@@ -1119,10 +1119,8 @@ mod tests {
let svc = SVC::fit(&x, &y, &params).unwrap();
// serialization
let _serialized_svc = &serde_json::to_string(&svc).unwrap();
let serialized_svc = &serde_json::to_string(&svc).unwrap();
// println!("{:?}", serialized_svc);
// TODO: for deserialization, deserialization is needed for `linalg::basic::matrix::DenseMatrix`
println!("{:?}", serialized_svc);
}
}