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

feat: documents matrix methods

Browse Source
This commit is contained in:
Volodymyr Orlov
2020-09-06 18:27:11 -07:00
parent 1e3ed4c924
commit bbe810d164
25 changed files with 587 additions and 245 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/linear/linear_regression.rs
+5 -5
View File
@@ -23,7 +23,7 @@
//! use smartcore::linear::linear_regression::*;
//!
//! // Longley dataset (https://www.statsmodels.org/stable/datasets/generated/longley.html)
//! let x = DenseMatrix::from_array(&[
//! let x = DenseMatrix::from_2d_array(&[
//! &[234.289, 235.6, 159.0, 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],
@@ -125,7 +125,7 @@ impl<T: RealNumber, M: Matrix<T>> LinearRegression<T, M> {
panic!("Number of rows of X doesn't match number of rows of Y");
}
let a = x.v_stack(&M::ones(x_nrows, 1));
let a = x.h_stack(&M::ones(x_nrows, 1));
let w = match parameters.solver {
LinearRegressionSolverName::QR => a.qr_solve_mut(b),
@@ -145,7 +145,7 @@ impl<T: RealNumber, M: Matrix<T>> LinearRegression<T, M> {
/// * `x` - _KxM_ data where _K_ is number of observations and _M_ is number of features.
pub fn predict(&self, x: &M) -> M::RowVector {
let (nrows, _) = x.shape();
let mut y_hat = x.dot(&self.coefficients);
let mut y_hat = x.matmul(&self.coefficients);
y_hat.add_mut(&M::fill(nrows, 1, self.intercept));
y_hat.transpose().to_row_vector()
}
@@ -168,7 +168,7 @@ mod tests {
#[test]
fn ols_fit_predict() {
let x = DenseMatrix::from_array(&[
let x = DenseMatrix::from_2d_array(&[
&[234.289, 235.6, 159.0, 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],
@@ -215,7 +215,7 @@ mod tests {
#[test]
fn serde() {
let x = DenseMatrix::from_array(&[
let x = DenseMatrix::from_2d_array(&[
&[234.289, 235.6, 159.0, 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],
src/linear/logistic_regression.rs
+18 -16
View File
@@ -14,7 +14,7 @@
//! use smartcore::linear::logistic_regression::*;
//!
//! //Iris data
//! let x = DenseMatrix::from_array(&[
//! 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],
@@ -277,8 +277,10 @@ impl<T: RealNumber, M: Matrix<T>> LogisticRegression<T, M> {
let mut result = M::zeros(1, n);
if self.num_classes == 2 {
let (nrows, _) = x.shape();
let x_and_bias = x.v_stack(&M::ones(nrows, 1));
let y_hat: Vec<T> = x_and_bias.dot(&self.weights.transpose()).to_raw_vector();
let x_and_bias = x.h_stack(&M::ones(nrows, 1));
let y_hat: Vec<T> = x_and_bias
.matmul(&self.weights.transpose())
.get_col_as_vec(0);
for i in 0..n {
result.set(
0,
@@ -288,8 +290,8 @@ impl<T: RealNumber, M: Matrix<T>> LogisticRegression<T, M> {
}
} else {
let (nrows, _) = x.shape();
let x_and_bias = x.v_stack(&M::ones(nrows, 1));
let y_hat = x_and_bias.dot(&self.weights.transpose());
let x_and_bias = x.h_stack(&M::ones(nrows, 1));
let y_hat = x_and_bias.matmul(&self.weights.transpose());
let class_idxs = y_hat.argmax();
for i in 0..n {
result.set(0, i, self.classes[class_idxs[i]]);
@@ -332,7 +334,7 @@ mod tests {
#[test]
fn multiclass_objective_f() {
let x = DenseMatrix::from_array(&[
let x = DenseMatrix::from_2d_array(&[
&[1., -5.],
&[2., 5.],
&[3., -2.],
@@ -363,16 +365,16 @@ mod tests {
objective.df(
&mut g,
&DenseMatrix::vector_from_array(&[1., 2., 3., 4., 5., 6., 7., 8., 9.]),
&DenseMatrix::row_vector_from_array(&[1., 2., 3., 4., 5., 6., 7., 8., 9.]),
);
objective.df(
&mut g,
&DenseMatrix::vector_from_array(&[1., 2., 3., 4., 5., 6., 7., 8., 9.]),
&DenseMatrix::row_vector_from_array(&[1., 2., 3., 4., 5., 6., 7., 8., 9.]),
);
assert!((g.get(0, 0) + 33.000068218163484).abs() < std::f64::EPSILON);
let f = objective.f(&DenseMatrix::vector_from_array(&[
let f = objective.f(&DenseMatrix::row_vector_from_array(&[
1., 2., 3., 4., 5., 6., 7., 8., 9.,
]));
@@ -381,7 +383,7 @@ mod tests {
#[test]
fn binary_objective_f() {
let x = DenseMatrix::from_array(&[
let x = DenseMatrix::from_2d_array(&[
&[1., -5.],
&[2., 5.],
&[3., -2.],
@@ -409,21 +411,21 @@ mod tests {
let mut g: DenseMatrix<f64> = DenseMatrix::zeros(1, 3);
objective.df(&mut g, &DenseMatrix::vector_from_array(&[1., 2., 3.]));
objective.df(&mut g, &DenseMatrix::vector_from_array(&[1., 2., 3.]));
objective.df(&mut g, &DenseMatrix::row_vector_from_array(&[1., 2., 3.]));
objective.df(&mut g, &DenseMatrix::row_vector_from_array(&[1., 2., 3.]));
assert!((g.get(0, 0) - 26.051064349381285).abs() < std::f64::EPSILON);
assert!((g.get(0, 1) - 10.239000702928523).abs() < std::f64::EPSILON);
assert!((g.get(0, 2) - 3.869294270156324).abs() < std::f64::EPSILON);
let f = objective.f(&DenseMatrix::vector_from_array(&[1., 2., 3.]));
let f = objective.f(&DenseMatrix::row_vector_from_array(&[1., 2., 3.]));
assert!((f - 59.76994756647412).abs() < std::f64::EPSILON);
}
#[test]
fn lr_fit_predict() {
let x = DenseMatrix::from_array(&[
let x = DenseMatrix::from_2d_array(&[
&[1., -5.],
&[2., 5.],
&[3., -2.],
@@ -460,7 +462,7 @@ mod tests {
#[test]
fn serde() {
let x = DenseMatrix::from_array(&[
let x = DenseMatrix::from_2d_array(&[
&[1., -5.],
&[2., 5.],
&[3., -2.],
@@ -489,7 +491,7 @@ mod tests {
#[test]
fn lr_fit_predict_iris() {
let x = DenseMatrix::from_array(&[
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],