pub fn make_hierarchical_clusters(
n_samples: usize,
n_features: usize,
n_main_clusters: usize,
n_sub_clusters: usize,
main_cluster_std: f64,
sub_cluster_std: f64,
randomseed: Option<u64>,
) -> Result<Dataset>Expand description
Generate hierarchical clusters (clusters within clusters)
Examples found in repository?
examples/complex_patterns_demo.rs (line 63)
12fn main() {
13 println!("=== Complex Pattern Generators Demonstration ===\n");
14
15 // Demonstrate non-linear pattern generators
16 println!("=== Non-Linear Pattern Generators ================");
17
18 // Spiral dataset
19 println!("1. Spiral Patterns:");
20 let spirals = make_spirals(200, 3, 0.1, Some(42)).unwrap();
21 println!(
22 " Generated {} spirals with {} samples",
23 3,
24 spirals.n_samples()
25 );
26 print_dataset_summary(&spirals, "Spirals");
27
28 // Two moons dataset
29 println!("\n2. Two Moons Pattern:");
30 let moons = make_moons(300, 0.05, Some(42)).unwrap();
31 print_dataset_summary(&moons, "Moons");
32
33 // Concentric circles
34 println!("\n3. Concentric Circles:");
35 let circles = make_circles(250, 0.4, 0.03, Some(42)).unwrap();
36 print_dataset_summary(&circles, "Circles");
37
38 // Swiss roll manifold
39 println!("\n4. Swiss Roll Manifold:");
40 let swiss_roll = make_swiss_roll(400, 0.1, Some(42)).unwrap();
41 print_dataset_summary(&swiss_roll, "Swiss Roll");
42 println!(" 3D manifold with intrinsic 2D structure");
43 println!();
44
45 // Demonstrate complex clustering patterns
46 println!("=== Complex Clustering Patterns ==================");
47
48 // Anisotropic (elongated) clusters
49 println!("1. Anisotropic Clusters:");
50 let aniso_clusters = make_anisotropic_blobs(300, 2, 4, 1.0, 5.0, Some(42)).unwrap();
51 print_dataset_summary(&aniso_clusters, "Anisotropic Clusters");
52 println!(" Elongated clusters with anisotropy factor 5.0");
53
54 // Different anisotropy factors demonstration
55 println!("\n Anisotropy Factor Comparison:");
56 for factor in [1.0, 2.0, 5.0, 10.0] {
57 let dataset = make_anisotropic_blobs(100, 2, 3, 1.0, factor, Some(42)).unwrap();
58 println!(" Factor {:.1}: {} clusters", factor, 3);
59 }
60
61 // Hierarchical clusters
62 println!("\n2. Hierarchical Clusters:");
63 let hierarchical = make_hierarchical_clusters(240, 3, 3, 4, 3.0, 0.8, Some(42)).unwrap();
64 print_dataset_summary(&hierarchical, "Hierarchical Clusters");
65 println!(" 3 main clusters, each with 4 sub-clusters");
66
67 if let Some(_metadata) = hierarchical.metadata.get("sub_cluster_labels") {
68 println!(" Sub-cluster structure preserved in metadata");
69 }
70 println!();
71
72 // Demonstrate parameter effects
73 println!("=== Parameter Effects Demonstration ==============");
74
75 // Noise effect on spirals
76 println!("1. Noise Effect on Spirals:");
77 for noise in [0.0, 0.05, 0.1, 0.2] {
78 let _spiraldata = make_spirals(100, 2, noise, Some(42)).unwrap();
79 println!(
80 " Noise {:.2}: Clean separation = {}",
81 noise,
82 if noise < 0.1 { "High" } else { "Low" }
83 );
84 }
85
86 // Factor effect on circles
87 println!("\n2. Factor Effect on Concentric Circles:");
88 for factor in [0.2, 0.4, 0.6, 0.8] {
89 let _circledata = make_circles(100, factor, 0.05, Some(42)).unwrap();
90 println!(" Factor {factor:.1}: Inner/Outer ratio = {factor:.1}");
91 }
92
93 // Cluster complexity in hierarchical patterns
94 println!("\n3. Hierarchical Cluster Complexity:");
95 for (main, sub) in [(2, 2), (2, 4), (3, 3), (4, 2)] {
96 let _hierdata = make_hierarchical_clusters(120, 2, main, sub, 2.0, 0.5, Some(42)).unwrap();
97 println!(
98 " {} main × {} sub = {} total clusters",
99 main,
100 sub,
101 main * sub
102 );
103 }
104 println!();
105
106 // Demonstrate use cases
107 println!("=== Use Cases and Applications ====================");
108
109 println!("1. **Non-linear Classification Testing**:");
110 println!(" - Spirals: Test algorithms that can handle multiple interleaved classes");
111 println!(" - Moons: Classic benchmark for non-linear separability");
112 println!(" - Circles: Test radial basis function methods");
113
114 println!("\n2. **Dimensionality Reduction Evaluation**:");
115 println!(" - Swiss Roll: Test manifold learning algorithms (t-SNE, UMAP, Isomap)");
116 println!(" - Preserves intrinsic 2D structure in 3D space");
117
118 println!("\n3. **Clustering Algorithm Testing**:");
119 println!(" - Anisotropic: Test algorithms robust to cluster shape variations");
120 println!(" - Hierarchical: Test multi-level clustering methods");
121
122 println!("\n4. **Robustness Testing**:");
123 println!(" - Variable noise levels test algorithm stability");
124 println!(" - Different cluster properties test generalization");
125 println!();
126
127 // Demonstrate advanced configurations
128 println!("=== Advanced Configuration Examples ===============");
129
130 println!("1. Multi-scale Spiral (Large dataset):");
131 let large_spirals = make_spirals(2000, 4, 0.08, Some(42)).unwrap();
132 print_dataset_summary(&large_spirals, "Large Spirals");
133
134 println!("\n2. High-dimensional Anisotropic Clusters:");
135 let hd_aniso = make_anisotropic_blobs(500, 10, 5, 1.5, 8.0, Some(42)).unwrap();
136 print_dataset_summary(&hd_aniso, "High-D Anisotropic");
137
138 println!("\n3. Deep Hierarchical Structure:");
139 let deep_hier = make_hierarchical_clusters(300, 4, 2, 6, 4.0, 1.0, Some(42)).unwrap();
140 print_dataset_summary(&deep_hier, "Deep Hierarchical");
141 println!(" Deep structure: 2 main → 12 sub-clusters");
142 println!();
143
144 // Performance and memory considerations
145 println!("=== Performance Guidelines =======================");
146 println!("**Recommended dataset sizes:**");
147 println!("- Development/Testing: 100-500 samples");
148 println!("- Algorithm benchmarking: 1,000-5,000 samples");
149 println!("- Performance testing: 10,000+ samples");
150
151 println!("\n**Memory usage (approximate):**");
152 println!("- Spirals (1000, 2D): ~16 KB");
153 println!("- Swiss Roll (1000, 3D): ~24 KB");
154 println!("- Hierarchical (1000, 5D): ~40 KB");
155
156 println!("\n**Parameter tuning tips:**");
157 println!("- Start with moderate noise (0.05-0.1)");
158 println!("- Use anisotropy factors 2.0-10.0 for clear elongation");
159 println!("- Keep sub-clusters ≤ 8 per main cluster for interpretability");
160 println!();
161
162 // Real-world applications
163 println!("=== Real-World Applications =======================");
164 println!("**Computer Vision:**");
165 println!("- Spirals: Object boundary detection");
166 println!("- Circles: Radial pattern recognition");
167
168 println!("\n**Machine Learning Research:**");
169 println!("- Benchmarking new clustering algorithms");
170 println!("- Testing manifold learning methods");
171 println!("- Evaluating non-linear classifiers");
172
173 println!("\n**Data Science Education:**");
174 println!("- Demonstrating algorithm limitations");
175 println!("- Visualizing high-dimensional data challenges");
176 println!("- Teaching feature engineering concepts");
177 println!();
178
179 println!("=== Complex Patterns Demo Complete ===============");
180}