Compare commits

...

2 Commits

4 changed files with 269 additions and 4 deletions

View File

@ -8,22 +8,55 @@ some basic inferential tests.
```rust ```rust
# extern crate rustframe; # extern crate rustframe;
use rustframe::compute::stats::{mean, mean_vertical, stddev, median}; use rustframe::compute::stats::{mean, mean_horizontal, mean_vertical, stddev, median, population_variance, percentile};
use rustframe::matrix::Matrix; use rustframe::matrix::Matrix;
let m = Matrix::from_vec(vec![1.0, 2.0, 3.0, 4.0], 2, 2); let m = Matrix::from_vec(vec![1.0, 2.0, 3.0, 4.0], 2, 2);
assert_eq!(mean(&m), 2.5); assert_eq!(mean(&m), 2.5);
assert_eq!(stddev(&m), 1.118033988749895); assert_eq!(stddev(&m), 1.118033988749895);
assert_eq!(median(&m), 2.5); assert_eq!(median(&m), 2.5);
assert_eq!(population_variance(&m), 1.25);
assert_eq!(percentile(&m, 50.0), 3.0);
// column averages returned as 1 x n matrix // column averages returned as 1 x n matrix
let row_means = mean_horizontal(&m);
assert_eq!(row_means.data(), &[2.0, 3.0]);
let col_means = mean_vertical(&m); let col_means = mean_vertical(&m);
assert_eq!(col_means.data(), &[1.5, 3.5]); assert_eq!(col_means.data(), & [1.5, 3.5]);
```
### Axis-specific Operations
Operations can be applied along specific axes (rows or columns):
```rust
# extern crate rustframe;
use rustframe::compute::stats::{mean_vertical, mean_horizontal, stddev_vertical, stddev_horizontal};
use rustframe::matrix::Matrix;
// 3x2 matrix
let m = Matrix::from_rows_vec(vec![1.0, 2.0, 3.0, 4.0, 5.0, 6.0], 3, 2);
// Mean along columns (vertical) - returns 1 x cols matrix
let col_means = mean_vertical(&m);
assert_eq!(col_means.shape(), (1, 2));
assert_eq!(col_means.data(), &[3.0, 4.0]); // [(1+3+5)/3, (2+4+6)/3]
// Mean along rows (horizontal) - returns rows x 1 matrix
let row_means = mean_horizontal(&m);
assert_eq!(row_means.shape(), (3, 1));
assert_eq!(row_means.data(), &[1.5, 3.5, 5.5]); // [(1+2)/2, (3+4)/2, (5+6)/2]
// Standard deviation along columns
let col_stddev = stddev_vertical(&m);
assert_eq!(col_stddev.shape(), (1, 2));
// Standard deviation along rows
let row_stddev = stddev_horizontal(&m);
assert_eq!(row_stddev.shape(), (3, 1));
``` ```
## Correlation ## Correlation
Correlation functions help measure linear relationships between datasets.
```rust ```rust
# extern crate rustframe; # extern crate rustframe;
use rustframe::compute::stats::{pearson, covariance}; use rustframe::compute::stats::{pearson, covariance};
@ -51,5 +84,68 @@ let pdf = normal_pdf(x, 0.0, 1.0);
assert_eq!(pdf.data().len(), 2); assert_eq!(pdf.data().len(), 2);
``` ```
### Additional Distributions
Rustframe provides several other probability distributions:
```rust
# extern crate rustframe;
use rustframe::compute::stats::distributions::{normal_cdf, binomial_pmf, binomial_cdf, poisson_pmf};
use rustframe::matrix::Matrix;
// Normal distribution CDF
let x = Matrix::from_vec(vec![0.0, 1.0], 1, 2);
let cdf = normal_cdf(x, 0.0, 1.0);
assert_eq!(cdf.data().len(), 2);
// Binomial distribution PMF
// Probability of k successes in n trials with probability p
let k = Matrix::from_vec(vec![0_u64, 1, 2, 3], 1, 4);
let pmf = binomial_pmf(3, k.clone(), 0.5);
assert_eq!(pmf.data().len(), 4);
// Binomial distribution CDF
let cdf = binomial_cdf(3, k, 0.5);
assert_eq!(cdf.data().len(), 4);
// Poisson distribution PMF
// Probability of k events with rate parameter lambda
let k = Matrix::from_vec(vec![0_u64, 1, 2], 1, 3);
let pmf = poisson_pmf(2.0, k);
assert_eq!(pmf.data().len(), 3);
```
### Inferential Statistics
Rustframe provides several inferential statistical tests:
```rust
# extern crate rustframe;
use rustframe::matrix::Matrix;
use rustframe::compute::stats::inferential::{t_test, chi2_test, anova};
// Two-sample t-test
let sample1 = Matrix::from_vec(vec![1.0, 2.0, 3.0, 4.0, 5.0], 1, 5);
let sample2 = Matrix::from_vec(vec![6.0, 7.0, 8.0, 9.0, 10.0], 1, 5);
let (t_statistic, p_value) = t_test(&sample1, &sample2);
assert!((t_statistic + 5.0).abs() < 1e-5);
assert!(p_value > 0.0 && p_value < 1.0);
// Chi-square test of independence
let observed = Matrix::from_vec(vec![12.0, 5.0, 8.0, 10.0], 2, 2);
let (chi2_statistic, p_value) = chi2_test(&observed);
assert!(chi2_statistic > 0.0);
assert!(p_value > 0.0 && p_value < 1.0);
// One-way ANOVA
let group1 = Matrix::from_vec(vec![1.0, 2.0, 3.0], 1, 3);
let group2 = Matrix::from_vec(vec![2.0, 3.0, 4.0], 1, 3);
let group3 = Matrix::from_vec(vec![3.0, 4.0, 5.0], 1, 3);
let groups = vec![&group1, &group2, &group3];
let (f_statistic, p_value) = anova(groups);
assert!(f_statistic > 0.0);
assert!(p_value > 0.0 && p_value < 1.0);
```
With the basics covered, explore predictive models in the With the basics covered, explore predictive models in the
[machine learning](./machine-learning.md) chapter. [machine learning](./machine-learning.md) chapter.

View File

@ -73,5 +73,85 @@ assert_eq!(frame.sum_vertical(), vec![3.0, 7.0]);
assert_eq!(frame.sum_horizontal(), vec![4.0, 6.0]); assert_eq!(frame.sum_horizontal(), vec![4.0, 6.0]);
``` ```
## Matrix Operations
```rust
# extern crate rustframe;
use rustframe::matrix::Matrix;
let data1 = Matrix::from_vec(vec![1.0, 2.0, 3.0, 4.0], 2, 2);
let data2 = Matrix::from_vec(vec![5.0, 6.0, 7.0, 8.0], 2, 2);
let sum = data1.clone() + data2.clone();
assert_eq!(sum.data(), vec![6.0, 8.0, 10.0, 12.0]);
let product = data1.clone() * data2.clone();
assert_eq!(product.data(), vec![5.0, 12.0, 21.0, 32.0]);
let scalar_product = data1.clone() * 2.0;
assert_eq!(scalar_product.data(), vec![2.0, 4.0, 6.0, 8.0]);
let equals = data1 == data1.clone();
assert_eq!(equals, true);
```
### Advanced Matrix Operations
Matrices support a variety of advanced operations:
```rust
# extern crate rustframe;
use rustframe::matrix::{Matrix, SeriesOps};
// Matrix multiplication (dot product)
let a = Matrix::from_vec(vec![1.0, 2.0, 3.0, 4.0], 2, 2);
let b = Matrix::from_vec(vec![5.0, 6.0, 7.0, 8.0], 2, 2);
let product = a.matrix_mul(&b);
assert_eq!(product.data(), vec![23.0, 34.0, 31.0, 46.0]);
// Transpose
let m = Matrix::from_vec(vec![1.0, 2.0, 3.0, 4.0], 2, 2);
let transposed = m.transpose();
assert_eq!(transposed.data(), vec![1.0, 3.0, 2.0, 4.0]);
// Map function over all elements
let m = Matrix::from_vec(vec![1.0, 2.0, 3.0, 4.0], 2, 2);
let squared = m.map(|x| x * x);
assert_eq!(squared.data(), vec![1.0, 4.0, 9.0, 16.0]);
// Zip two matrices with a function
let a = Matrix::from_vec(vec![1.0, 2.0, 3.0, 4.0], 2, 2);
let b = Matrix::from_vec(vec![5.0, 6.0, 7.0, 8.0], 2, 2);
let zipped = a.zip(&b, |x, y| x + y);
assert_eq!(zipped.data(), vec![6.0, 8.0, 10.0, 12.0]);
```
### Matrix Reductions
Matrices support various reduction operations:
```rust
# extern crate rustframe;
use rustframe::matrix::{Matrix, SeriesOps};
let m = Matrix::from_rows_vec(vec![1.0, 2.0, 3.0, 4.0, 5.0, 6.0], 3, 2);
// Sum along columns (vertical)
let col_sums = m.sum_vertical();
assert_eq!(col_sums, vec![9.0, 12.0]); // [1+3+5, 2+4+6]
// Sum along rows (horizontal)
let row_sums = m.sum_horizontal();
assert_eq!(row_sums, vec![3.0, 7.0, 11.0]); // [1+2, 3+4, 5+6]
// Cumulative sum along columns
let col_cumsum = m.cumsum_vertical();
assert_eq!(col_cumsum.data(), vec![1.0, 4.0, 9.0, 2.0, 6.0, 12.0]);
// Cumulative sum along rows
let row_cumsum = m.cumsum_horizontal();
assert_eq!(row_cumsum.data(), vec![1.0, 3.0, 5.0, 3.0, 7.0, 11.0]);
```
With the basics covered, continue to the [compute features](./compute.md) With the basics covered, continue to the [compute features](./compute.md)
chapter for statistics and analytics. chapter for statistics and analytics.

View File

@ -72,5 +72,75 @@ let transformed = pca.transform(&data);
assert_eq!(transformed.cols(), 1); assert_eq!(transformed.cols(), 1);
``` ```
### Gaussian Naive Bayes
Gaussian Naive Bayes classifier for continuous features:
```rust
# extern crate rustframe;
use rustframe::compute::models::gaussian_nb::GaussianNB;
use rustframe::matrix::Matrix;
// Training data with 2 features
let x = Matrix::from_rows_vec(vec![
1.0, 2.0,
2.0, 3.0,
3.0, 4.0,
4.0, 5.0
], 4, 2);
// Class labels (0 or 1)
let y = Matrix::from_vec(vec![0.0, 0.0, 1.0, 1.0], 4, 1);
// Train the model
let mut model = GaussianNB::new(1e-9, true);
model.fit(&x, &y);
// Make predictions
let predictions = model.predict(&x);
assert_eq!(predictions.rows(), 4);
```
### Dense Neural Networks
Simple fully connected neural network:
```rust
# extern crate rustframe;
use rustframe::compute::models::dense_nn::{DenseNN, DenseNNConfig, ActivationKind, InitializerKind, LossKind};
use rustframe::matrix::Matrix;
// Training data with 2 features
let x = Matrix::from_rows_vec(vec![
0.0, 0.0,
0.0, 1.0,
1.0, 0.0,
1.0, 1.0
], 4, 2);
// XOR target outputs
let y = Matrix::from_vec(vec![0.0, 1.0, 1.0, 0.0], 4, 1);
// Create a neural network with 2 hidden layers
let config = DenseNNConfig {
input_size: 2,
hidden_layers: vec![4, 4],
output_size: 1,
activations: vec![ActivationKind::Sigmoid, ActivationKind::Sigmoid, ActivationKind::Sigmoid],
initializer: InitializerKind::Uniform(0.5),
loss: LossKind::MSE,
learning_rate: 0.1,
epochs: 1000,
};
let mut model = DenseNN::new(config);
// Train the model
model.train(&x, &y);
// Make predictions
let predictions = model.predict(&x);
assert_eq!(predictions.rows(), 4);
```
For helper functions and upcoming modules, visit the For helper functions and upcoming modules, visit the
[utilities](./utilities.md) section. [utilities](./utilities.md) section.

View File

@ -35,6 +35,25 @@ let v2 = rng.next_u64();
assert_ne!(v1, v2); assert_ne!(v1, v2);
``` ```
## Stats Functions
```rust
# extern crate rustframe;
use rustframe::matrix::Matrix;
use rustframe::compute::stats::descriptive::{mean, median, stddev};
let data = Matrix::from_vec(vec![1.0, 2.0, 3.0, 4.0, 5.0], 1, 5);
let mean_value = mean(&data);
assert_eq!(mean_value, 3.0);
let median_value = median(&data);
assert_eq!(median_value, 3.0);
let std_value = stddev(&data);
assert_eq!(std_value, 2.0_f64.sqrt());
```
Upcoming utilities will cover: Upcoming utilities will cover:
- Data import/export helpers - Data import/export helpers