NNTrainedModel.java

/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
package org.apache.tika.detect;

public class NNTrainedModel extends TrainedModel {

    private final int numOfInputs;
    private final int numOfHidden;
    private final int numOfOutputs;
    private final float[][] Theta1;
    private final float[][] Theta2;

    public NNTrainedModel(final int nInput, final int nHidden, final int nOutput,
                          final float[] nn_params) {
        this.numOfInputs = nInput;
        this.numOfHidden = nHidden;
        this.numOfOutputs = nOutput;
        this.Theta1 = new float[numOfHidden][numOfInputs + 1];
        this.Theta2 = new float[numOfOutputs][numOfHidden + 1];
        populateThetas(nn_params);
    }

    // convert the vector params to the 2 thetas.
    private void populateThetas(final float[] nn_params) {
        int m = this.Theta1.length;
        int n = this.Theta1[0].length;
        int i, j, k = 0;
        for (i = 0; i < n; i++) {
            for (j = 0; j < m; j++) {
                this.Theta1[j][i] = nn_params[k];
                k++;
            }
        }

        m = this.Theta2.length;
        n = this.Theta2[0].length;
        for (i = 0; i < n; i++) {
            for (j = 0; j < m; j++) {
                this.Theta2[j][i] = nn_params[k];
                k++;
            }
        }
    }

    @Override
    public double predict(double[] unseen) {
        // TODO Auto-generated method stub
        return 0;
    }

    /**
     * The given input vector of unseen is m=(256 + 1) * n= 1 this returns a
     * prediction probability
     */
    @Override
    public float predict(float[] unseen) {
        // please ensure the unseen in size consistent with theta1

        int i, j;
        int m = this.Theta1.length;
        int n = this.Theta1[0].length;
        float[] hh = new float[m + 1];// hidden unit summation
        hh[0] = 1;
        for (i = 0; i < m; i++) {
            double h = 0;
            for (j = 0; j < n; j++) {
                h += this.Theta1[i][j] * unseen[j];
            }
            // apply sigmoid
            h = 1.0 / (1.0 + Math.exp(-h));
            hh[i + 1] = (float) h;
        }

        m = this.Theta2.length;
        n = this.Theta2[0].length;
        float[] oo = new float[m];
        for (i = 0; i < m; i++) {
            double o = 0;
            for (j = 0; j < n; j++) {
                o += this.Theta2[i][j] * hh[j];
            }
            // apply sigmoid
            o = 1.0 / (1.0 + Math.exp(-o));
            oo[i] = (float) o;
        }

        return oo[0];
    }
}