向量余弦值(Cosine)(C#)

时间：2011-04-11 来源：live41

接前面一篇TF-IDF的代码,用其结果来计算出向量的余弦值的代码

using System;
using System.Collections.Generic;
using System.Text;

namespace Cluster
{
    static class Cosine
    {
        /// <summary>
        /// 计算向量余弦值
        /// </summary>
        /// <param name="vector1"></param>
        /// <param name="vector2"></param>
        public static double Calculate(Dictionary<int, double> vector1, Dictionary<int, double> vector2)
        {
            double dotProduct = CalcDotProduct(vector1, vector2);
            double length1 = CalcLength(vector1);
            double length2 = CalcLength(vector2);
            double cosine = dotProduct / (length1 * length2);

            return cosine;
        }

        /// <summary>
        /// 计算向量长度(vector length)
        /// </summary>
        /// <param name="vector"></param>
        /// <returns></returns>
        private static double CalcLength(Dictionary<int, double> vector)
        {
            double length = 0;
            foreach (KeyValuePair<int, double> kvp in vector)
            {
                length += Math.Pow(kvp.Value, 2);
            }
            return Math.Sqrt(length);
        }

        /// <summary>
        /// 计算向量点积(dot product)/内积(inner product)
        /// </summary>
        /// <param name="vector1"></param>
        /// <param name="vector2"></param>
        /// <returns></returns>
        private static double CalcDotProduct(Dictionary<int, double> vector1, Dictionary<int, double> vector2)
        {
            double dotProduct = 0;
            foreach (KeyValuePair<int, double> kvp in vector1)
            {
                if (vector2.ContainsKey(kvp.Key))
                {
                    dotProduct += kvp.Value * vector2[kvp.Key];
                }
            }
            return dotProduct;
        }
    }
}