Linear model train params change to enum. see #71

alibaba · Apr 9, 2020 · db3fd54 · db3fd54
1 parent aa15681
commit db3fd54
Show file tree

Hide file tree

Showing 17 changed files with 88 additions and 103 deletions.
diff --git a/.../src/main/java/com/alibaba/alink/operator/batch/classification/LinearSvmTrainBatchOp.java b/.../src/main/java/com/alibaba/alink/operator/batch/classification/LinearSvmTrainBatchOp.java
@@ -3,7 +3,6 @@
 import com.alibaba.alink.operator.common.linear.BaseLinearModelTrainBatchOp;
 import com.alibaba.alink.operator.common.linear.LinearModelType;
 import com.alibaba.alink.params.classification.LinearBinaryClassTrainParams;
-import com.alibaba.alink.params.classification.LinearSvmTrainParams;
 
 import org.apache.flink.ml.api.misc.param.Params;
 
@@ -13,14 +12,13 @@
  *
  */
 public final class LinearSvmTrainBatchOp extends BaseLinearModelTrainBatchOp<LinearSvmTrainBatchOp>
-	implements LinearSvmTrainParams <LinearSvmTrainBatchOp> {
+	implements LinearBinaryClassTrainParams<LinearSvmTrainBatchOp> {
 
 	public LinearSvmTrainBatchOp() {
 		this(new Params());
 	}
 
 	public LinearSvmTrainBatchOp(Params params) {
-		super(params.set(LinearBinaryClassTrainParams.L_2, 1.0 / params.get(LinearSvmTrainParams.C)),
-			LinearModelType.SVM, "Linear SVM");
+		super(params, LinearModelType.SVM, "Linear SVM");
 	}
 }
diff --git a/.../java/com/alibaba/alink/operator/batch/classification/LogisticRegressionTrainBatchOp.java b/.../java/com/alibaba/alink/operator/batch/classification/LogisticRegressionTrainBatchOp.java
@@ -2,7 +2,7 @@
 
 import com.alibaba.alink.operator.common.linear.BaseLinearModelTrainBatchOp;
 import com.alibaba.alink.operator.common.linear.LinearModelType;
-import com.alibaba.alink.params.classification.LogisticRegressionTrainParams;
+import com.alibaba.alink.params.classification.LinearBinaryClassTrainParams;
 
 import org.apache.flink.ml.api.misc.param.Params;
 
@@ -12,7 +12,7 @@
  *
  */
 public final class LogisticRegressionTrainBatchOp extends BaseLinearModelTrainBatchOp<LogisticRegressionTrainBatchOp>
-	implements LogisticRegressionTrainParams <LogisticRegressionTrainBatchOp> {
+	implements LinearBinaryClassTrainParams<LogisticRegressionTrainBatchOp> {
 
 	public LogisticRegressionTrainBatchOp() {
 		this(new Params());

diff --git a/core/src/main/java/com/alibaba/alink/operator/batch/classification/SoftmaxTrainBatchOp.java b/core/src/main/java/com/alibaba/alink/operator/batch/classification/SoftmaxTrainBatchOp.java
@@ -17,7 +17,6 @@
 import com.alibaba.alink.operator.common.linear.LinearModelDataConverter;
 import com.alibaba.alink.operator.common.linear.SoftmaxObjFunc;
 import com.alibaba.alink.operator.common.optim.Lbfgs;
-import com.alibaba.alink.operator.common.optim.OptimMethod;
 import com.alibaba.alink.operator.common.optim.OptimizerFactory;
 import com.alibaba.alink.operator.common.optim.Owlqn;
 import com.alibaba.alink.operator.common.optim.objfunc.OptimObjFunc;
@@ -212,8 +211,8 @@ public void reduce(Iterable<Integer> values, Collector<OptimObjFunc> out) throws
         });
 
         // solve the opt problem.
-        if (params.contains("optimMethod")) {
-            OptimMethod method = OptimMethod.valueOf(params.get(LinearTrainParams.OPTIM_METHOD).toUpperCase());
+        if (params.contains(LinearTrainParams.OPTIM_METHOD)) {
+            OptimMethod method = params.get(LinearTrainParams.OPTIM_METHOD);
             return OptimizerFactory.create(objFunc, trainData, coefDim, params, method)
                 .optimize();
         } else if (params.get(SoftmaxTrainParams.L_1) > 0) {

diff --git a/core/src/main/java/com/alibaba/alink/operator/common/linear/BaseLinearModelTrainBatchOp.java b/core/src/main/java/com/alibaba/alink/operator/common/linear/BaseLinearModelTrainBatchOp.java
@@ -11,7 +11,6 @@
 import com.alibaba.alink.operator.batch.BatchOperator;
 import com.alibaba.alink.operator.common.linear.unarylossfunc.*;
 import com.alibaba.alink.operator.common.optim.Lbfgs;
-import com.alibaba.alink.operator.common.optim.OptimMethod;
 import com.alibaba.alink.operator.common.optim.OptimizerFactory;
 import com.alibaba.alink.operator.common.optim.Owlqn;
 import com.alibaba.alink.operator.common.optim.objfunc.OptimObjFunc;
@@ -258,7 +257,7 @@ public static DataSet<Tuple2<DenseVector, double[]>> optimize(Params params,
             .fromElements(getObjFunction(modelType, params));
 
         if (params.contains(LinearTrainParams.OPTIM_METHOD)) {
-            OptimMethod method = OptimMethod.valueOf(params.get(LinearTrainParams.OPTIM_METHOD).toUpperCase());
+            LinearTrainParams.OptimMethod method = params.get(LinearTrainParams.OPTIM_METHOD);
             return OptimizerFactory.create(objFunc, trainData, coefficientDim, params, method).optimize();
         } else if (params.get(HasL1.L_1) > 0) {
             return new Owlqn(objFunc, trainData, coefficientDim, params).optimize();

diff --git a/core/src/main/java/com/alibaba/alink/operator/common/optim/Gd.java b/core/src/main/java/com/alibaba/alink/operator/common/optim/Gd.java
@@ -18,6 +18,7 @@
 import com.alibaba.alink.common.comqueue.IterativeComQueue;
 import com.alibaba.alink.common.comqueue.communication.AllReduce;
 import com.alibaba.alink.params.shared.iter.HasMaxIterDefaultAs100;
+import com.alibaba.alink.params.shared.linear.LinearTrainParams.OptimMethod;
 import com.alibaba.alink.params.shared.optim.HasNumSearchStepDv4;
 
 import org.apache.flink.api.java.DataSet;

diff --git a/core/src/main/java/com/alibaba/alink/operator/common/optim/Lbfgs.java b/core/src/main/java/com/alibaba/alink/operator/common/optim/Lbfgs.java
@@ -19,6 +19,7 @@
 import com.alibaba.alink.common.comqueue.IterativeComQueue;
 import com.alibaba.alink.common.comqueue.communication.AllReduce;
 import com.alibaba.alink.params.shared.iter.HasMaxIterDefaultAs100;
+import com.alibaba.alink.params.shared.linear.LinearTrainParams.OptimMethod;
 import com.alibaba.alink.params.shared.optim.HasNumSearchStepDv4;
 
 import org.apache.flink.api.java.DataSet;

diff --git a/core/src/main/java/com/alibaba/alink/operator/common/optim/OptimMethod.java b/core/src/main/java/com/alibaba/alink/operator/common/optim/OptimMethod.java
diff --git a/core/src/main/java/com/alibaba/alink/operator/common/optim/OptimizerFactory.java b/core/src/main/java/com/alibaba/alink/operator/common/optim/OptimizerFactory.java
@@ -2,6 +2,7 @@
 
 import com.alibaba.alink.common.linalg.Vector;
 import com.alibaba.alink.operator.common.optim.objfunc.OptimObjFunc;
+import com.alibaba.alink.params.shared.linear.LinearTrainParams;
 
 import org.apache.flink.api.java.DataSet;
 import org.apache.flink.api.java.tuple.Tuple3;
@@ -16,11 +17,11 @@ public static Optimizer create(
 		DataSet <Tuple3 <Double, Double, Vector>> trainData,
 		DataSet <Integer> coefDim,
 		Params params,
-		OptimMethod method) {
+		LinearTrainParams.OptimMethod method) {
 		switch (method) {
 			case SGD:
 				return new Sgd(objFunc, trainData, coefDim, params);
-			case NEWTON:
+			case Newton:
 				return new Newton(objFunc, trainData, coefDim, params);
 			case GD:
 				return new Gd(objFunc, trainData, coefDim, params);

diff --git a/core/src/main/java/com/alibaba/alink/operator/common/optim/Owlqn.java b/core/src/main/java/com/alibaba/alink/operator/common/optim/Owlqn.java
@@ -21,6 +21,7 @@
 
 import com.alibaba.alink.params.shared.linear.HasL1;
 import com.alibaba.alink.params.shared.iter.HasMaxIterDefaultAs100;
+import com.alibaba.alink.params.shared.linear.LinearTrainParams.OptimMethod;
 import com.alibaba.alink.params.shared.optim.HasNumSearchStepDv4;
 
 import org.apache.flink.api.java.DataSet;

diff --git a/core/src/main/java/com/alibaba/alink/operator/common/optim/Sgd.java b/core/src/main/java/com/alibaba/alink/operator/common/optim/Sgd.java
@@ -18,6 +18,7 @@
 import com.alibaba.alink.common.comqueue.ComputeFunction;
 import com.alibaba.alink.common.comqueue.IterativeComQueue;
 import com.alibaba.alink.common.comqueue.communication.AllReduce;
+import com.alibaba.alink.params.shared.linear.LinearTrainParams.OptimMethod;
 import com.alibaba.alink.params.shared.optim.SgdParams;
 
 import org.apache.flink.api.java.DataSet;

diff --git a/core/src/main/java/com/alibaba/alink/operator/common/optim/subfunc/CalcLosses.java b/core/src/main/java/com/alibaba/alink/operator/common/optim/subfunc/CalcLosses.java
@@ -4,10 +4,10 @@
 
 import com.alibaba.alink.common.linalg.DenseVector;
 import com.alibaba.alink.common.linalg.Vector;
-import com.alibaba.alink.operator.common.optim.OptimMethod;
 import com.alibaba.alink.operator.common.optim.objfunc.OptimObjFunc;
 import com.alibaba.alink.common.comqueue.ComContext;
 import com.alibaba.alink.common.comqueue.ComputeFunction;
+import com.alibaba.alink.params.shared.linear.LinearTrainParams.OptimMethod;
 
 import org.apache.flink.api.java.tuple.Tuple2;
 import org.apache.flink.api.java.tuple.Tuple3;

diff --git a/core/src/main/java/com/alibaba/alink/operator/common/optim/subfunc/UpdateModel.java b/core/src/main/java/com/alibaba/alink/operator/common/optim/subfunc/UpdateModel.java
@@ -1,11 +1,11 @@
 package com.alibaba.alink.operator.common.optim.subfunc;
 
 import com.alibaba.alink.common.linalg.DenseVector;
-import com.alibaba.alink.operator.common.optim.OptimMethod;
 import com.alibaba.alink.common.comqueue.ComContext;
 import com.alibaba.alink.common.comqueue.ComputeFunction;
 import com.alibaba.alink.params.shared.iter.HasMaxIterDefaultAs100;
 import com.alibaba.alink.params.shared.linear.HasEpsilonDv0000001;
+import com.alibaba.alink.params.shared.linear.LinearTrainParams.OptimMethod;
 
 import org.apache.flink.api.java.tuple.Tuple2;
 import org.apache.flink.ml.api.misc.param.Params;

diff --git a/core/src/main/java/com/alibaba/alink/params/classification/LinearSvmTrainParams.java b/core/src/main/java/com/alibaba/alink/params/classification/LinearSvmTrainParams.java
diff --git a/.../src/main/java/com/alibaba/alink/params/classification/LogisticRegressionTrainParams.java b/.../src/main/java/com/alibaba/alink/params/classification/LogisticRegressionTrainParams.java
diff --git a/core/src/main/java/com/alibaba/alink/params/shared/linear/LinearTrainParams.java b/core/src/main/java/com/alibaba/alink/params/shared/linear/LinearTrainParams.java
@@ -1,5 +1,6 @@
 package com.alibaba.alink.params.shared.linear;
 
+import com.alibaba.alink.params.ParamUtil;
 import com.alibaba.alink.params.shared.colname.HasFeatureColsDefaultAsNull;
 import com.alibaba.alink.params.shared.colname.HasLabelCol;
 import com.alibaba.alink.params.shared.colname.HasVectorColDefaultAsNull;
@@ -11,29 +12,58 @@
 
 /**
  * parameters of linear training.
- *
  */
 public interface LinearTrainParams<T> extends
-	HasWithIntercept <T>,
-	HasMaxIterDefaultAs100<T>,
-	HasEpsilonDv0000001 <T>,
-	HasFeatureColsDefaultAsNull <T>,
-	HasLabelCol <T>,
-	HasWeightColDefaultAsNull <T>,
-	HasVectorColDefaultAsNull <T>,
-	HasStandardization <T> {
-
-	ParamInfo<String> OPTIM_METHOD = ParamInfoFactory
-		.createParamInfo("optimMethod", String.class)
-		.setDescription("optimization method")
-		.setHasDefaultValue(null)
-		.build();
-
-	default String getOptimMethod() {
-		return get(OPTIM_METHOD);
-	}
-
-	default T setOptimMethod(String value) {
-		return set(OPTIM_METHOD, value);
-	}
+    HasWithIntercept<T>,
+    HasMaxIterDefaultAs100<T>,
+    HasEpsilonDv0000001<T>,
+    HasFeatureColsDefaultAsNull<T>,
+    HasLabelCol<T>,
+    HasWeightColDefaultAsNull<T>,
+    HasVectorColDefaultAsNull<T>,
+    HasStandardization<T> {
+
+    ParamInfo<OptimMethod> OPTIM_METHOD = ParamInfoFactory
+        .createParamInfo("optimMethod", OptimMethod.class)
+        .setDescription("optimization method")
+        .setHasDefaultValue(null)
+        .build();
+
+    default OptimMethod getOptimMethod() {
+        return get(OPTIM_METHOD);
+    }
+
+    default T setOptimMethod(String value) {
+        return set(OPTIM_METHOD, ParamUtil.searchEnum(OPTIM_METHOD, value));
+    }
+
+    default T setOptimMethod(OptimMethod value) {
+        return set(OPTIM_METHOD, value);
+    }
+
+    /**
+     * Optimization Type.
+     */
+    enum OptimMethod {
+        /**
+         * LBFGS method
+         */
+        LBFGS,
+        /**
+         * Gradient Descent method
+         */
+        GD,
+        /**
+         * Newton method
+         */
+        Newton,
+        /**
+         * Stochastic Gradient Descent method
+         */
+        SGD,
+        /**
+         * OWLQN method
+         */
+        OWLQN
+    }
 }
diff --git a/core/src/main/java/com/alibaba/alink/pipeline/classification/LinearSvm.java b/core/src/main/java/com/alibaba/alink/pipeline/classification/LinearSvm.java
@@ -2,29 +2,30 @@
 
 import com.alibaba.alink.operator.batch.BatchOperator;
 import com.alibaba.alink.operator.batch.classification.LinearSvmTrainBatchOp;
+
 import org.apache.flink.ml.api.misc.param.Params;
+
+import com.alibaba.alink.params.classification.LinearBinaryClassTrainParams;
 import com.alibaba.alink.params.classification.LinearSvmPredictParams;
-import com.alibaba.alink.params.classification.LinearSvmTrainParams;
 import com.alibaba.alink.pipeline.Trainer;
 
 /**
  * Linear svm pipeline op.
- *
  */
-public class LinearSvm extends Trainer <LinearSvm, LinearSvmModel>
-	implements LinearSvmTrainParams <LinearSvm>, LinearSvmPredictParams <LinearSvm> {
+public class LinearSvm extends Trainer<LinearSvm, LinearSvmModel>
+    implements LinearBinaryClassTrainParams<LinearSvm>, LinearSvmPredictParams<LinearSvm> {
 
-	public LinearSvm() {
-		super();
-	}
+    public LinearSvm() {
+        super();
+    }
 
-	public LinearSvm(Params params) {
-		super(params);
-	}
+    public LinearSvm(Params params) {
+        super(params);
+    }
 
-	@Override
-	protected BatchOperator train(BatchOperator in) {
-		return new LinearSvmTrainBatchOp(this.getParams()).linkFrom(in);
-	}
+    @Override
+    protected BatchOperator train(BatchOperator in) {
+        return new LinearSvmTrainBatchOp(this.getParams()).linkFrom(in);
+    }
 
 }
diff --git a/core/src/main/java/com/alibaba/alink/pipeline/classification/LogisticRegression.java b/core/src/main/java/com/alibaba/alink/pipeline/classification/LogisticRegression.java
@@ -3,17 +3,18 @@
 import com.alibaba.alink.operator.batch.BatchOperator;
 import com.alibaba.alink.operator.batch.classification.LogisticRegressionTrainBatchOp;
 import org.apache.flink.ml.api.misc.param.Params;
+
+import com.alibaba.alink.params.classification.LinearBinaryClassTrainParams;
 import com.alibaba.alink.params.classification.LogisticRegressionPredictParams;
-import com.alibaba.alink.params.classification.LogisticRegressionTrainParams;
 import com.alibaba.alink.pipeline.Trainer;
 
 /**
  * Logistic regression is a popular method to predict a categorical response.
  *
  */
 public class LogisticRegression extends Trainer <LogisticRegression, LogisticRegressionModel> implements
-	LogisticRegressionTrainParams <LogisticRegression>,
-	LogisticRegressionPredictParams <LogisticRegression> {
+    LogisticRegressionPredictParams <LogisticRegression>,
+	LinearBinaryClassTrainParams<LogisticRegression> {
 
 	public LogisticRegression() {super();}