Wsrtjtyk

I have some vectors of data as:

681,305,305,304,667,92.87,254.51,100.24  has "A" as key

and

460,682,702,683,443,89.17,273.83,102.19 has"B" as key

and

593,395,413,418,564,71.13,17.67,83.79 has "C" as key

Then, from my device, I obtained a sample of data and needed to translate it into key. sample of data as below:

593,395,413,418,564,71.13,17.67,83.79

I want to calculate the similarity of the obtained data with the three samples of the above data, see which one is most similar to the obtained data, then I choose key of sample to return. I'm using C#.

Is there any algorithm for calculating it? Please help me, thanks!

There are many possible approaches:

1. Calculate Minkowski distances between possible pairs and choose the closest.




2. For every two arrays A and B, you can calculate an array of differences/deltas D, where D_i = |A_i - B_i|. For this array you can calculate mean, variance, maybe other moments and infer from it.

...

But how you will use these quantities depends on the meaning of your data and on how it will impact performance.

As you pointed out in the comment, the data actually contains two parts: forces and angles, so most likely you need to compare not the whole arrays, but two subarrays of lengths 4 and 3.

There can be several assumptions based upon that you may make decision for similarity. I have presented a simple program in C# for finding similarity based upon the summation of absolute difference between values in same indices. The one with minimum summation will be most similar. You may extend your solution in similar manner.

using System;

using System.Collections.Generic;

public class Test

{

  public static void Main()

  {

     Dictionary<string, List<double> > map = new Dictionary<string, List<double> >();

     List<double>listA =  new List<double>(new double{681,305,305,304,667,92.87,254.51,100.24});

     map.Add("A", listA);

     List<double>listB =  new List<double>(new double{460,682,702,683,443,89.17,273.83,102.19});

     map.Add("B", listB);

     List<double>listC = new List<double>(new double{593,395,413,418,564,71.13,17.67,83.79});

     map.Add("C", listC);

     List<double>testList = new List<double>(new double{593,395,413,418,564,71.13,17.67,83.79});

     string ans = mostSimilarKey(testList, map);

     Console.WriteLine("Most similar key is: "+ans);

  }



  static string mostSimilarKey(List<double>testList, Dictionary<string, List<double> > map) {

    double minDifference = Double.MaxValue;

    string ans = "";

    foreach (var pair in map) {

        double absoluteDifference = getAbsoluteDifference(testList, pair.Value);

        if (absoluteDifference < minDifference) {

            minDifference = absoluteDifference;

            ans = pair.Key;

        }

    }

    return ans;

 }



 static double getAbsoluteDifference(List<double>testList, List<double>list) {

    double absoluteDifference = 0.0;

    for (int i = 0; i < testList.Count; ++i) {

        absoluteDifference += Math.Abs(testList[i] - list[i]);

    }

    return absoluteDifference;

  }

}