Added support for the AR-only ARIMA model type

README.md

@@ -20,6 +20,8 @@ Java library and command-line application for converting [StatsModels](https://w
     * [Poisson](https://www.statsmodels.org/dev/generated/statsmodels.discrete.discrete_model.Poisson.html)
     * [OrderedModel](https://www.statsmodels.org/dev/generated/statsmodels.miscmodels.ordinal_model.OrderedModel.html):
       * Distributions: `logit`, `probit`
+  * Univariate Time-Series Analysis:
+    * [ARIMA](https://www.statsmodels.org/dev/generated/statsmodels.tsa.arima.model.ARIMA.html)
 * Production quality:
   * Complete test coverage.
   * Fully compliant with the [JPMML-Evaluator](https://github.com/jpmml/jpmml-evaluator) library.

pmml-statsmodels/src/main/java/statsmodels/tools/Bunch.java

@@ -0,0 +1,32 @@
+/*
+ * Copyright (c) 2024 Villu Ruusmann
+ *
+ * This file is part of JPMML-StatsModels
+ *
+ * JPMML-StatsModels is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * JPMML-StatsModels is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with JPMML-StatsModels.  If not, see <http://www.gnu.org/licenses/>.
+ */
+package statsmodels.tools;
+
+import org.jpmml.python.PythonObject;
+
+public class Bunch extends PythonObject {
+
+	public Bunch(String module, String name){
+		super(module, name);
+	}
+
+	public void __setstate__(Bunch bunch){
+		super.__setstate__(bunch);
+	}
+}

pmml-statsmodels/src/main/java/statsmodels/tsa/arima/ARIMA.java

@@ -0,0 +1,145 @@
+/*
+ * Copyright (c) 2024 Villu Ruusmann
+ *
+ * This file is part of JPMML-StatsModels
+ *
+ * JPMML-StatsModels is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * JPMML-StatsModels is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with JPMML-StatsModels.  If not, see <http://www.gnu.org/licenses/>.
+ */
+package statsmodels.tsa.arima;
+
+import java.util.List;
+
+import org.dmg.pmml.Array;
+import org.dmg.pmml.DataField;
+import org.dmg.pmml.DataType;
+import org.dmg.pmml.MiningField.UsageType;
+import org.dmg.pmml.MiningFunction;
+import org.dmg.pmml.MiningSchema;
+import org.dmg.pmml.OpType;
+import org.dmg.pmml.time_series.InterceptVector;
+import org.dmg.pmml.time_series.MeasurementMatrix;
+import org.dmg.pmml.time_series.StateSpaceModel;
+import org.dmg.pmml.time_series.StateVector;
+import org.dmg.pmml.time_series.TransitionMatrix;
+import org.jpmml.converter.CMatrix;
+import org.jpmml.converter.Matrix;
+import org.jpmml.converter.ModelUtil;
+import org.jpmml.converter.PMMLUtil;
+import org.jpmml.converter.Schema;
+import org.jpmml.python.HasArray;
+import org.jpmml.statsmodels.StatsModelsEncoder;
+import statsmodels.Results;
+import statsmodels.tsa.TimeSeriesModel;
+import statsmodels.tsa.statespace.SmootherResults;
+
+public class ARIMA extends TimeSeriesModel {
+
+	public ARIMA(String module, String name){
+		super(module, name);
+	}
+
+	@Override
+	public Schema encodeSchema(StatsModelsEncoder encoder){
+		Schema schema = super.encodeSchema(encoder);
+
+		@SuppressWarnings("unused")
+		DataField dataField = encoder.createDataField("horizon", OpType.CONTINUOUS, DataType.INTEGER);
+
+		return schema;
+	}
+
+	@Override
+	public org.dmg.pmml.time_series.TimeSeriesModel encodeModel(Results results, Schema schema){
+		HasArray predictedState = results.getArray("predicted_state");
+		SmootherResults smootherResults = results.get("smoother_results", SmootherResults.class);
+
+		HasArray design = smootherResults.getDesign();
+		HasArray obsIntercept = smootherResults.getObsIntercept();
+		HasArray transition = smootherResults.getTransition();
+
+		MiningSchema miningSchema = ModelUtil.createMiningSchema(schema.getLabel())
+			.addMiningFields(ModelUtil.createMiningField("horizon", UsageType.SUPPLEMENTARY));
+
+		StateVector stateVector = new StateVector(createRealArray(predictedState, -1));
+
+		MeasurementMatrix measurementMatrix = new MeasurementMatrix(createMatrix(design));
+
+		TransitionMatrix transitionMatrix = new TransitionMatrix(createMatrix(transition));
+
+		InterceptVector interceptVector = new InterceptVector(createRealArray(obsIntercept, -1))
+			.setType(InterceptVector.Type.OBSERVATION);
+
+		StateSpaceModel stateSpaceModel = new StateSpaceModel()
+			.setStateVector(stateVector)
+			.setMeasurementMatrix(measurementMatrix)
+			.setTransitionMatrix(transitionMatrix)
+			.setInterceptVector(interceptVector);
+
+		org.dmg.pmml.time_series.TimeSeriesModel timeSeriesModel = new org.dmg.pmml.time_series.TimeSeriesModel(MiningFunction.TIME_SERIES, org.dmg.pmml.time_series.TimeSeriesModel.Algorithm.STATE_SPACE_MODEL, miningSchema)
+			.setStateSpaceModel(stateSpaceModel);
+
+		return timeSeriesModel;
+	}
+
+	static
+	private Array createRealArray(HasArray hasArray, int column){
+		Matrix<?> matrix = toMatrix(hasArray);
+
+		List<? extends Number> columnValues;
+
+		if(column >= 0){
+			columnValues = (List)matrix.getColumnValues(column);
+		} else
+
+		{
+			columnValues = (List)matrix.getColumnValues(matrix.getColumns() + column);
+		}
+
+		return PMMLUtil.createRealArray(columnValues);
+	}
+
+	static
+	private org.dmg.pmml.Matrix createMatrix(HasArray hasArray){
+		Matrix<?> matrix = toMatrix(hasArray);
+
+		org.dmg.pmml.Matrix result = new org.dmg.pmml.Matrix()
+			.setNbRows(matrix.getRows())
+			.setNbCols(matrix.getColumns());
+
+		for(int row = 0; row < matrix.getRows(); row++){
+			List<? extends Number> rowValues = (List)matrix.getRowValues(row);
+
+			result.addArrays(PMMLUtil.createRealArray(rowValues));
+		}
+
+		return result;
+	}
+
+	static
+	private Matrix<?> toMatrix(HasArray hasArray){
+		int[] shape = hasArray.getArrayShape();
+		List<?> values = hasArray.getArrayContent();
+
+		if(shape.length == 3){
+
+			if(shape[2] != 1){
+				throw new IllegalArgumentException();
+			}
+		}
+
+		return new CMatrix(values, shape[0], shape[1]);
+	}
+
+	private static final String NAME_INDEX = "index";
+}

pmml-statsmodels/src/main/java/statsmodels/tsa/statespace/Initialization.java

@@ -0,0 +1,40 @@
+/*
+ * Copyright (c) 2024 Villu Ruusmann
+ *
+ * This file is part of JPMML-StatsModels
+ *
+ * JPMML-StatsModels is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * JPMML-StatsModels is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with JPMML-StatsModels.  If not, see <http://www.gnu.org/licenses/>.
+ */
+package statsmodels.tsa.statespace;
+
+import org.jpmml.python.CythonObject;
+
+public class Initialization extends CythonObject {
+
+	public Initialization(String module, String name){
+		super(module, name);
+	}
+
+	@Override
+	public void __init__(Object[] args){
+		super.__setstate__(INIT_ATTRIBUTES, args);
+	}
+
+	private static final String[] INIT_ATTRIBUTES = {
+		"k_states",
+		"constant",
+		"stationary_cov",
+		"approximate_diffuse_variance"
+	};
+}

pmml-statsmodels/src/main/java/statsmodels/tsa/statespace/KalmanFilter.java

@@ -0,0 +1,44 @@
+/*
+ * Copyright (c) 2024 Villu Ruusmann
+ *
+ * This file is part of JPMML-StatsModels
+ *
+ * JPMML-StatsModels is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * JPMML-StatsModels is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with JPMML-StatsModels.  If not, see <http://www.gnu.org/licenses/>.
+ */
+package statsmodels.tsa.statespace;
+
+import org.jpmml.python.CythonObject;
+
+public class KalmanFilter extends CythonObject {
+
+	public KalmanFilter(String module, String name){
+		super(module, name);
+	}
+
+	@Override
+	public void __init__(Object[] args){
+		super.__setstate__(INIT_ATTRIBUTES, args);
+	}
+
+	private static final String[] INIT_ATTRIBUTES = {
+		"model",
+		"filter_method",
+		"inversion_method",
+		"stability_method",
+		"conserve_memory",
+		"filter_timing",
+		"tolerance",
+		"loglikelihood_burn"
+	};
+}

pmml-statsmodels/src/main/java/statsmodels/tsa/statespace/KalmanSmoother.java

@@ -0,0 +1,40 @@
+/*
+ * Copyright (c) 2024 Villu Ruusmann
+ *
+ * This file is part of JPMML-StatsModels
+ *
+ * JPMML-StatsModels is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * JPMML-StatsModels is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with JPMML-StatsModels.  If not, see <http://www.gnu.org/licenses/>.
+ */
+package statsmodels.tsa.statespace;
+
+import org.jpmml.python.CythonObject;
+
+public class KalmanSmoother extends CythonObject {
+
+	public KalmanSmoother(String module, String name){
+		super(module, name);
+	}
+
+	@Override
+	public void __init__(Object[] args){
+		super.__setstate__(INIT_ATTRIBUTES, args);
+	}
+
+	private static final String[] INIT_ATTRIBUTES = {
+		"model",
+		"kfilter",
+		"smoother_output",
+		"smooth_method"
+	};
+}

pmml-statsmodels/src/main/java/statsmodels/tsa/statespace/SmootherResults.java

@@ -0,0 +1,41 @@
+/*
+ * Copyright (c) 2024 Villu Ruusmann
+ *
+ * This file is part of JPMML-StatsModels
+ *
+ * JPMML-StatsModels is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * JPMML-StatsModels is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with JPMML-StatsModels.  If not, see <http://www.gnu.org/licenses/>.
+ */
+package statsmodels.tsa.statespace;
+
+import org.jpmml.python.HasArray;
+import org.jpmml.python.PythonObject;
+
+public class SmootherResults extends PythonObject {
+
+	public SmootherResults(String module, String name){
+		super(module, name);
+	}
+
+	public HasArray getDesign(){
+		return getArray("design");
+	}
+
+	public HasArray getObsIntercept(){
+		return getArray("obs_intercept");
+	}
+
+	public HasArray getTransition(){
+		return getArray("transition");
+	}
+}