Dataframe driven API for predictors

CHANGELOG.md

@@ -1,5 +1,9 @@
 # Changelog
 
+## 13.0.0
+- Predictor's API: `DataFrame` used instead of `Matrix`
+- `DecisionTreeSolver`: data splitting logic fixed
+
 ## 12.1.2
 - `xrange` package version locked
 

benchmark/cross_validator.dart

@@ -1,19 +1,16 @@
 // 8.5 sec
-import 'dart:async';
-
 import 'package:benchmark_harness/benchmark_harness.dart';
 import 'package:ml_algo/ml_algo.dart';
+import 'package:ml_dataframe/ml_dataframe.dart';
 import 'package:ml_linalg/matrix.dart';
 import 'package:ml_linalg/vector.dart';
 
 const observationsNum = 1000;
-const featuresNum = 20;
+const columnsNum = 21;
 
 class CrossValidatorBenchmark extends BenchmarkBase {
   CrossValidatorBenchmark() : super('Cross validator benchmark');
 
-  Matrix features;
-  Matrix labels;
   CrossValidator crossValidator;
 
   static void main() {
@@ -22,18 +19,20 @@ class CrossValidatorBenchmark extends BenchmarkBase {
 
   @override
   void run() {
-    crossValidator.evaluate((trainFeatures, trainLabels) =>
-        ParameterlessRegressor.knn(trainFeatures, trainLabels, k: 7),
-        features, labels, MetricType.mape);
+    crossValidator.evaluate((trainSamples, targetFeatureNames) =>
+        KnnRegressor(trainSamples, targetFeatureNames.first, k: 7),
+        MetricType.mape);
   }
 
   @override
   void setup() {
-    features = Matrix.fromRows(List.generate(observationsNum,
-            (i) => Vector.randomFilled(featuresNum)));
-    labels = Matrix.fromColumns([Vector.randomFilled(observationsNum)]);
+    final samples = Matrix.fromRows(List.generate(observationsNum,
+            (i) => Vector.randomFilled(columnsNum)));
+
+    final dataFrame = DataFrame.fromMatrix(samples);
 
-    crossValidator = CrossValidator.kFold(numberOfFolds: 5);
+    crossValidator = CrossValidator.kFold(dataFrame, ['col_20'],
+        numberOfFolds: 5);
   }
 
   void tearDown() {}

benchmark/knn_regression.dart → benchmark/knn_regressor.dart

@@ -1,22 +1,22 @@
-// 5.7 sec
-import 'dart:async';
-
+// 10.0 sec (MacBook Air mid 2017)
 import 'package:benchmark_harness/benchmark_harness.dart';
 import 'package:ml_algo/ml_algo.dart';
+import 'package:ml_algo/src/regressor/knn_regressor_impl.dart';
+import 'package:ml_dataframe/ml_dataframe.dart';
 import 'package:ml_linalg/matrix.dart';
 import 'package:ml_linalg/vector.dart';
 
 const observationsNum = 500;
 const featuresNum = 20;
 
 class KnnRegressorBenchmark extends BenchmarkBase {
-  KnnRegressorBenchmark() : super('KNN regression benchmark');
+  KnnRegressorBenchmark() : super('Knn regression benchmark');
 
   Matrix features;
-  Matrix testFeatures;
+  DataFrame testFeatures;
   Matrix labels;
   Matrix testLabels;
-  ParameterlessRegressor regressor;
+  KnnRegressor regressor;
 
 
   static void main() {
@@ -34,11 +34,17 @@ class KnnRegressorBenchmark extends BenchmarkBase {
             (i) => Vector.randomFilled(featuresNum)));
     labels = Matrix.fromColumns([Vector.randomFilled(observationsNum * 2)]);
 
-    testFeatures = Matrix.fromRows(List.generate(observationsNum,
-            (i) => Vector.randomFilled(featuresNum)));
+    testFeatures = DataFrame.fromMatrix(
+        Matrix.fromRows(
+            List.generate(
+                observationsNum,
+                (i) => Vector.randomFilled(featuresNum),
+            ),
+        ),
+    );
     testLabels = Matrix.fromColumns([Vector.randomFilled(observationsNum)]);
 
-    regressor = ParameterlessRegressor.knn(features, labels, k: 7);
+    regressor = KnnRegressorImpl(features, labels, 'target', k: 7);
   }
 
   void tearDown() {}

benchmark/linear_regressor.dart

@@ -0,0 +1,44 @@
+import 'package:benchmark_harness/benchmark_harness.dart';
+import 'package:ml_algo/ml_algo.dart';
+import 'package:ml_dataframe/ml_dataframe.dart';
+import 'package:ml_linalg/matrix.dart';
+import 'package:ml_linalg/vector.dart';
+
+const observationsNum = 200;
+const featuresNum = 20;
+
+class LinearRegressorBenchmark extends BenchmarkBase {
+  LinearRegressorBenchmark() : super('Linear regressor');
+
+  DataFrame fittingData;
+
+  static void main() {
+    LinearRegressorBenchmark().report();
+  }
+
+  @override
+  void run() {
+    LinearRegressor(fittingData, 'col_20');
+  }
+
+  @override
+  void setup() {
+    final features = Matrix.fromRows(List.generate(observationsNum,
+            (i) => Vector.randomFilled(featuresNum)));
+
+    final labels = Matrix.fromColumns([Vector.randomFilled(observationsNum)]);
+
+    fittingData = DataFrame.fromMatrix(
+        Matrix.fromColumns([
+          ...features.columns,
+          ...labels.columns,
+        ]),
+    );
+  }
+
+  void tearDown() {}
+}
+
+Future main() async {
+  LinearRegressorBenchmark.main();
+}

benchmark/logistic_regressor.dart

@@ -0,0 +1,42 @@
+import 'package:benchmark_harness/benchmark_harness.dart';
+import 'package:ml_algo/ml_algo.dart';
+import 'package:ml_dataframe/ml_dataframe.dart';
+import 'package:ml_linalg/matrix.dart';
+import 'package:ml_linalg/vector.dart';
+
+const observationsNum = 200;
+const columnsNum = 21;
+
+class LogisticRegressorBenchmark extends BenchmarkBase {
+  LogisticRegressorBenchmark() : super('Logistic regressor');
+
+  DataFrame _data;
+
+  static void main() {
+    LogisticRegressorBenchmark().report();
+  }
+
+  @override
+  void run() {
+    LogisticRegressor(
+      _data,
+      'col_20',
+      minCoefficientsUpdate: null,
+      iterationsLimit: 200,
+    );
+  }
+
+  @override
+  void setup() {
+    final Matrix observations = Matrix.fromRows(List.generate(observationsNum,
+            (i) => Vector.randomFilled(columnsNum)));
+
+    _data = DataFrame.fromMatrix(observations);
+  }
+
+  void tearDown() {}
+}
+
+Future main() async {
+  LogisticRegressorBenchmark.main();
+}

benchmark/main.dart

@@ -1,12 +1,10 @@
-import 'dart:async';
-
-import 'gradient_descent_regression.dart' as gradientDescentRegressionBenchmark;
-import 'logistic_regression.dart' as logisticRegressionBenchmark;
-import 'algorithms/knn.dart' as knnBenchmark;
+import 'linear_regressor.dart' as gradient_descent_regression_benchmark;
+import 'logistic_regressor.dart' as logistic_regression_benchmark;
+import 'algorithms/knn.dart' as knn_regressor_benchmark;
 
 Future main() async {
   //  (MacBook Air mid 2017)
-  await gradientDescentRegressionBenchmark.main(); // 0.07 sec
-  await logisticRegressionBenchmark.main(); // 0.12 sec
-  await knnBenchmark.main(); // 5 sec
+  await gradient_descent_regression_benchmark.main(); // 0.07 sec
+  await logistic_regression_benchmark.main(); // 0.12 sec
+  await knn_regressor_benchmark.main(); // 5 sec
 }

example/main.dart

@@ -1,29 +1,20 @@
-import 'dart:async';
-
 import 'package:ml_algo/ml_algo.dart';
-import 'package:ml_linalg/matrix.dart';
+import 'package:ml_dataframe/ml_dataframe.dart';
 
 /// A simple usage example using synthetic data. To see more complex examples,
 /// please, visit other directories in this folder
 Future main() async {
-  // Let's create a feature matrix (a set of independent variables)
-  final features = Matrix.fromList([
-    [2.0, 3.0, 4.0, 5.0],
-    [12.0, 32.0, 1.0, 3.0],
-    [27.0, 3.0, 0.0, 59.0],
-  ]);
-
-  // Let's create dependent variables vector. It will be used as `true` values
-  // to adjust regression coefficients
-  final labels = Matrix.fromList([
-    [4.3],
-    [3.5],
-    [2.1],
-  ]);
+  // Let's create a dataframe with fitting data, let's assume, that the target
+  // column is the fifth column (column with index 4)
+  final dataFrame = DataFrame(<Iterable<num>>[
+    [ 2,  3, 4,  5, 4.3],
+    [12, 32, 1,  3, 3.5],
+    [27,  3, 0, 59, 2.1],
+  ], headerExists: false);
 
   // Let's create a regressor itself and train it
-  final regressor = LinearRegressor.gradient(
-      features, labels,
+  final regressor = LinearRegressor(
+      dataFrame, 'col_4',
       iterationsLimit: 100,
       initialLearningRate: 0.0005,
       learningRateType: LearningRateType.constant);

lib/ml_algo.dart

@@ -1,10 +1,13 @@
 export 'package:ml_algo/src/algorithms/knn/kernel_type.dart';
-export 'package:ml_algo/src/classifier/linear/logistic_regressor/logistic_regressor.dart';
-export 'package:ml_algo/src/classifier/linear/softmax_regressor/softmax_regressor.dart';
+export 'package:ml_algo/src/classifier/decision_tree_classifier.dart';
+export 'package:ml_algo/src/classifier/logistic_regressor.dart';
+export 'package:ml_algo/src/classifier/softmax_regressor.dart';
+export 'package:ml_algo/src/linear_optimizer/gradient_optimizer/learning_rate_generator/learning_rate_type.dart';
+export 'package:ml_algo/src/linear_optimizer/linear_optimizer_type.dart';
+export 'package:ml_algo/src/linear_optimizer/regularization_type.dart';
 export 'package:ml_algo/src/metric/classification/type.dart';
 export 'package:ml_algo/src/metric/metric_type.dart';
 export 'package:ml_algo/src/metric/regression/type.dart';
 export 'package:ml_algo/src/model_selection/cross_validator/cross_validator.dart';
+export 'package:ml_algo/src/regressor/knn_regressor.dart';
 export 'package:ml_algo/src/regressor/linear_regressor.dart';
-export 'package:ml_algo/src/regressor/parameterless_regressor.dart';
-export 'package:ml_algo/src/solver/linear/gradient/learning_rate_generator/learning_rate_type.dart';