C# Forecast 预测数据方法，MathNet.Numerics学习分享

news2024/10/6 14:32:27

Forecast 数据预测

        /// <summary>
        /// 数据预测
        /// </summary>
        /// <param name="xValues">数据序列</param>
        /// <param name="yValues">数据值</param>
        /// <param name="forecastPoint">第N个预测序列</param>
        /// <returns>返回第N个序列预测值</returns>
        public double Forecast(double[] xValues, double[] yValues, double forecastPoint)
        {
            //数组平均值
            var xAverage = xValues.Average();
            var yAverage = yValues.Average();

            //预测值上下区间
            var bounds = yValues
                .Select((y, i) => new { Value = y, Index = i })
                .Aggregate(new { Top = 0.0, Bottom = 0.0 }, (acc, cur) =>
                    new
                    {
                        //上限值
                        Top = acc.Top + (xValues[cur.Index] - xAverage) * (yValues[cur.Index] - yAverage),
                        //下限值
                        Bottom = acc.Bottom + Math.Pow(xValues[cur.Index] - xAverage, 2.0)
                    });
            
            var level = bounds.Top / bounds.Bottom;
            //区中预测值
            return (yAverage - level * xAverage) + level * forecastPoint;
        }

测试调用：

        //测试数据
            double[] xList = new double[] { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 };
            double[] yList = new double[] { 836.43, 831.68, 836.03, 835.94, 836.61, 835.8, 835.2, 834.35, 835.93, 835.41, 835.75, 838.15, 835.19, 836.64, 746.17, 669.6, 666.57, 679.41, 679.99, 684.96 };

            //根据已知20个值预测未来10个值
            for (int i = 1; i <= 10; i++)
            {
                //调用主方法
                double foreCastValue = Forecast(xList, yList, 20 + i);
                //打印输出
                Response.Write(foreCastValue.ToString("F4") + "\r\n");
            }