Esempio n. 1
0
        public void Test()
        {
            int testSize = 10;

            for (int i = 0; i < testSize; i++)
            {
                int      index      = _rnd.Next(_dataMgr.count - 1);
                double[] testInput  = _dataMgr.GetInputData(index);
                double   testOutput = _dataMgr.GetLabelData(index);

                double predictOuput = _regressionMachine.Predict(testInput);

                Console.WriteLine("Test Index:" + index + ", Expected:" + testOutput + " Predict:" + predictOuput);
            }

            double mse = 0.0;

            for (int i = 0; i < _output.Length; i++)
            {
                double[] testInput    = _dataMgr.GetInputData(i);
                double   testOutput   = _dataMgr.GetLabelData(i);
                double   predictOuput = _regressionMachine.Predict(testInput);

                mse += (testOutput - predictOuput) * (testOutput - predictOuput);
            }

            mse = Math.Sqrt(mse) / _output.Length;

            Console.WriteLine("MSE:" + mse);
        }
Esempio n. 2
0
        public KernelRegressionTester(HouseDataMgr dataMgr, KernelRegression machine = null)
        {
            _dataMgr = dataMgr;

            _input  = new double[_dataMgr.count][];
            _output = new double[_dataMgr.count];

            for (int i = 0; i < _dataMgr.count; i++)
            {
                _input[i]  = _dataMgr.GetInputData(i);
                _output[i] = _dataMgr.GetLabelData(i);
            }

            if (null == machine)
            {
                _regressionMachine = new KernelRegression();
            }
            else
            {
                _regressionMachine = machine;
            }

            _rnd = new Random();
        }