Encoding the label column of a file

[george-vendras]

Hi there. I am new to github. Today, I would try to encode the label column (of type string) into digits. Since I am not strong in c# prog, I could not quite handle IEnumerators and all these info I would find here and there, thus I was somewhat forced to write my own code, a process that took me around 5 hours 'cause C# debugger would ask me this or that all the time. I do not intend in hot encoding.
The idea here is that the label column is at the end of the matrix, perhaps, I could had added an if-statement to check whether the label column is number type or string type, okay, of course I have sense that my code is not efficient. that is because I am here to ask you:

1. Technical advice on how should I improve code with various c# cookies such as IEnumerator or whatever else,
2. What else would you like to see in the code. I mean, would you like to see a complete lib-like? And what functionalities should it contain (like hotencoding etc)? "Future directions" in a few words. I do not know, there are a lot of cases and scenarios.
3. Using MathNet library, I tried to find a function to perform a dot product but it was not possible. Can somebody tell me how I can perform a dotproduct using MathNet?

A few times I see codes that create an object in order, for example, to calculate the determinant and I wonder why should this be the case and not be hosted into Program class as a internal or private static function.

Thanx in advance!

P.S. I have not done research for similar discussion because I searched a lot in google but returns were somewhat poor.

using System;
using System.Text;
using static System.Console;
using System.Globalization;

namespace SupportVectormachinePerceptron
{
class Program
{
public static void Main(string[] args)
{
Console.Title = "Support Vector Machine Perceptron Calculation";

        string filenamePath = string.Empty;
        List<List<decimal>> records = null;

        //Perceptron's Utilities
        decimal result = 0, compare = 0, bias_b = 0;
        decimal[] Weights, X; 
        bool misclassified = true;


        Write("Please, enter the full path of the data file you want to import: ");
        filenamePath = Console.ReadLine();// C:\Users\YourName\Desktop\irisDataset.txt

        records = ReadCSVFile(filenamePath);

        Console.WriteLine("records type: {0}", records.GetType());

        foreach (List<decimal> record in records)
        {
            foreach (decimal item in record)
            {
                Console.Write(item + "\t");
            }
            Console.WriteLine();
        }

        for (int i = 0; i < records.Count; i++)
        {
            for (int j = 0; j < records[i].Count; j++)
                Console.Write(records[i][j] + "\t");
            Console.WriteLine();
        }


        // - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -


        //Perceptron Calculation
        Weights = new decimal[records[0].Count];//5th elements of matrix Weights is dedicated to bias b calculation
        X = new decimal[records[0].Count];

        while (misclassified)
        {
            misclassified = false;
            for(int i = 0; i < records.Count; i++) 
            {
                for (int j = 0; j < records[i].Count - 1; j++)
                    X[j] = records[i][j];
                X[records[i].Count - 1] = 1;
                
                IEnumerable<decimal> dotProducts = Weights.Zip(X, (a, b) => a * b);
                foreach (decimal dp in dotProducts)
                { 
                    result += dp;
                }                    
                
                compare = records[i][records[i].Count - 1] * result;
                result = 0;

                if (compare < 1)
                {                        
                    for (int j = 0; j < Weights.Length; j++)
                        Weights[j] += records[i][records[i].Count - 1] * X[j];

                    misclassified = true;
                }
            }
        }

        //Separate bias b from Weights matrix
        bias_b = Weights[Weights.Length - 1];
        Weights = Weights.Where((e, i) => i != Weights.Length - 1).ToArray();

        Console.WriteLine("\n\nFor your dataset, the Support Vector Machine weights matrix is the following:");
        foreach(var w in Weights)
            Console.Write(w + "\t");
        Console.WriteLine("\nBias b equals to: {0}", bias_b);

        //End of Perceptron program
        Console.WriteLine("\n\n* * End of Perceptron calculation. * *");
    }
    private static List<List<decimal>> ReadCSVFile(string filenamePath)
    {
        //Variable definitions
        int i = 0, j = 0;
        string[] readAllLines, strNums, columnEncodement;

        List<string> columnEncodementList;
        List<List<decimal>> Dataset;


        //Variable initializations            
        readAllLines = File.ReadAllLines(filenamePath);
        columnEncodement = new string[readAllLines.Length];

        columnEncodementList = new List<string>();
        Dataset = new List<List<decimal>>();

        //Instantiate the 2D List and add necessary number of inner Lists innerList as rows
        for (int k = 0; k < readAllLines.Length; k++)
        {
            var innerList = new List<decimal>();
            Dataset.Add(innerList);
        }

        var uniqueItemsList = readAllLines.Distinct().ToList();

        foreach (var uI in uniqueItemsList)
        {
            strNums = uI.Split(',');
            columnEncodementList.Add(strNums[strNums.Length - 1]);
            i++;
        }
        i = 0;

        columnEncodement = columnEncodementList.Distinct().ToArray();
        columnEncodementList.Clear();//No need of columnEncodementList anymore

        foreach (string str in readAllLines)
        {
            //Add number columns to List
            strNums = str.Split(',');
            for (j = 0; j < strNums.Length - 1; j++)
                Dataset[i].Add(Convert.ToDecimal(strNums[j], CultureInfo.InvariantCulture.NumberFormat));
            
            //Encode Label column into -1 and 1 and add to the irisDataset list
            if (strNums[j] == columnEncodement[0])
                Dataset[i].Add(-1);
            else
                Dataset[i].Add(1);

            j = 0;
            i++;
        }
        return Dataset;
    }
}

}

[george-vendras]

For example, in

var uniqueItemsList = readAllLines.Distinct().ToList();

foreach (var uI in uniqueItemsList)
{
    strNums = uI.Split(',');
    columnEncodementList.Add(strNums[strNums.Length - 1]);
    i++;
}

how could I had avoided "uniqueItemsList"?