data_generators/
data_generators.rs

1use scirs2_datasets::{
2    make_blobs, make_classification, make_regression, make_time_series, utils::normalize,
3    utils::train_test_split,
4};
5
6fn main() -> Result<(), Box<dyn std::error::Error>> {
7    println!("Creating synthetic datasets...\n");
8
9    // Generate classification dataset
10    let n_samples = 100;
11    let n_features = 5;
12
13    let classification_data = make_classification(
14        n_samples,
15        n_features,
16        3,        // 3 classes
17        2,        // 2 clusters per class
18        3,        // 3 informative features
19        Some(42), // random seed
20    )?;
21
22    // Train-test split
23    let (train, test) = train_test_split(&classification_data, 0.2, Some(42))?;
24
25    println!("Classification dataset:");
26    println!("  Total samples: {}", classification_data.n_samples());
27    println!("  Features: {}", classification_data.n_features());
28    println!("  Training samples: {}", train.n_samples());
29    println!("  Test samples: {}", test.n_samples());
30
31    // Generate regression dataset
32    let regression_data = make_regression(
33        n_samples,
34        n_features,
35        3,   // 3 informative features
36        0.5, // noise level
37        Some(42),
38    )?;
39
40    println!("\nRegression dataset:");
41    println!("  Samples: {}", regression_data.n_samples());
42    println!("  Features: {}", regression_data.n_features());
43
44    // Normalize the data (in-place)
45    let mut data_copy = regression_data.data.clone();
46    normalize(&mut data_copy);
47    println!("  Data normalized successfully");
48
49    // Generate clustering data (blobs)
50    let clustering_data = make_blobs(
51        n_samples,
52        2,   // 2 features for easy visualization
53        4,   // 4 clusters
54        0.8, // cluster standard deviation
55        Some(42),
56    )?;
57
58    println!("\nClustering dataset (blobs):");
59    println!("  Samples: {}", clustering_data.n_samples());
60    println!("  Features: {}", clustering_data.n_features());
61
62    // Find the number of clusters by finding the max value of target
63    let num_clusters = clustering_data.target.as_ref().map_or(0, |t| {
64        let mut max_val = -1.0;
65        for &val in t.iter() {
66            if val > max_val {
67                max_val = val;
68            }
69        }
70        (max_val as usize) + 1
71    });
72
73    println!("  Clusters: {}", num_clusters);
74
75    // Generate time series data
76    let time_series = make_time_series(
77        100,  // 100 time steps
78        3,    // 3 features/variables
79        true, // with trend
80        true, // with seasonality
81        0.2,  // noise level
82        Some(42),
83    )?;
84
85    println!("\nTime series dataset:");
86    println!("  Time steps: {}", time_series.n_samples());
87    println!("  Features: {}", time_series.n_features());
88
89    Ok(())
90}
data_generators/data_generators.rs

data_generators/
data_generators.rs
