Update readme with classification run instructions

Author: MarcoDiFrancesco

.gitignore

@@ -6,4 +6,5 @@ target/
 /.vscode/
 # Local configuration
 .cargo/config.toml
-/.venv*/
+/.venv*/
+generate_data_synthetic.py

README.md

@@ -43,7 +43,9 @@ cargo run --release --example credit_card
 
 ### 📊 Classification
 
-🏗️ We plan to implement Aggregated Mondrian Forests.
+```sh
+RUSTFLAGS=-Awarnings cargo run --release --example synthetic
+```
 
 ### 🛒 Recsys
 

examples/classification/synthetic.rs

@@ -34,7 +34,6 @@ fn get_labels(transactions: IterCsv<f32, File>) -> Vec<String> {
 }
 
 fn main() {
-    let now = Instant::now();
     let n_trees: usize = 1;
 
     let transactions_f = Synthetic::load_data();
@@ -47,6 +46,8 @@ fn main() {
         MondrianForestClassifier::new(n_trees, features.len(), labels.len());
     let mut score_total = 0.0;
 
+    let now = Instant::now();
+
     let transactions = Synthetic::load_data();
     for (idx, transaction) in transactions.enumerate() {
         let data = transaction.unwrap();
@@ -71,13 +72,19 @@ fn main() {
             // println!("=M=3 score: {:?}", score);
             score_total += score;
 
+            // println!(
+            //     "{score_total} / {idx} = {}",
+            //     score_total / idx.to_f32().unwrap()
+            // );
+        }
+        if idx == 100_000 - 1 {
             println!(
-                "{score_total} / {idx} = {}",
+                "Accuracy: {score_total} / {idx} = {}",
                 score_total / idx.to_f32().unwrap()
             );
         }
 
-        println!("=M=1 partial_fit {x_ord}");
+        // println!("=M=1 partial_fit {x_ord}");
         mf.partial_fit(&x_ord, y);
     }
 

generate_data.py

@@ -185,7 +185,7 @@ impl<F: FType> MondrianTreeClassifier<F> {
             let exp_dist = Exp::new(lambda.to_f32().unwrap()).unwrap();
             let exp_sample = F::from_f32(exp_dist.sample(&mut self.rng)).unwrap();
             // DEBUG: shadowing with Exp expected value
-            let exp_sample = F::one() / lambda;
+            // let exp_sample = F::one() / lambda;
             exp_sample
         };
         let split_time = self.compute_split_time(time, exp_sample, node_idx, y, extensions.sum());
@@ -202,7 +202,7 @@ impl<F: FType> MondrianTreeClassifier<F> {
                     .collect::<Array1<F>>();
                 let e_sample = F::from_f32(self.rng.gen::<f32>()).unwrap() * extensions.sum();
                 // DEBUG: shadowing with expected value
-                let e_sample = F::from_f32(0.5).unwrap() * extensions.sum();
+                // let e_sample = F::from_f32(0.5).unwrap() * extensions.sum();
                 cumsum.iter().position(|&val| val > e_sample).unwrap()
             };
 
@@ -219,7 +219,7 @@ impl<F: FType> MondrianTreeClassifier<F> {
             };
             let threshold = F::from_f32(self.rng.gen_range(lower_bound..upper_bound)).unwrap();
             // DEBUG: split in the middle
-            let threshold = F::from_f32((lower_bound + upper_bound) / 2.0).unwrap();
+            // let threshold = F::from_f32((lower_bound + upper_bound) / 2.0).unwrap();
 
             let mut min_list = node_min_list.clone();
             let mut max_list = node_max_list.clone();