form1.cs

It s the kNN or else named k-Nearest Neabours algorythm, realized by the tools of Visual Studio deve

using System;
using System.Collections.Generic;
using System.ComponentModel;
using System.Data;
using System.Drawing;
using System.Text;
using System.Windows.Forms;
using System.Collections;
using System.IO;

namespace kNN
{
    public partial class Form1 : Form
    {

        private DataTable table = new DataTable();
        private DataTable teachingTable = new DataTable();
        private DataTable testTable = new DataTable();
        private DataTable classifiedTable = new DataTable();
        private ArrayList TargetAttr = new ArrayList();

        private ArrayList trainingSamples = new ArrayList();
        private ArrayList testSamples = new ArrayList();
        private ArrayList answers = new ArrayList();
        
        
        public Form1()
        {
            InitializeComponent();
        }

        public void OpenSample()
        {
            this.table.Clear();
            this.table.Rows.Clear();
            this.table.Columns.Clear();
            this.comboBox1.Items.Clear();

            if (openFileDialog1.ShowDialog() == DialogResult.OK)
            {
                StreamReader reader = new StreamReader(openFileDialog1.FileName);
                string[] attributes = reader.ReadLine().Split(',');

                DataRow row = this.table.NewRow();
                for (int i = 0; i < attributes.Length; i++)
                //foreach (string value in attributes)
                {

                    this.table.Columns.Add();
                    this.teachingTable.Columns.Add();
                    this.testTable.Columns.Add();
                    row[i] = attributes[i];
                    this.table.Columns[i].ColumnName = ("Attribute" + (i + 1).ToString());

                }
                // default target attribute is the  last one
                int j = 0;
                foreach (string value in attributes)
                {
                    comboBox1.Items.Add("Attribute " + (++j).ToString());

                }
                comboBox1.SelectedIndex = comboBox1.Items.Count - 1;

                this.table.Rows.Add(row);

                string line;

                while ((line = reader.ReadLine()) != null)
                {
                    row = this.table.NewRow();
                    string[] sample = line.Split(',');
                    for (int i = 0; i < sample.Length; i++)
                    {
                        row[i] = sample[i];
                    }
                    this.table.Rows.Add(row);
                }
            }
        }


        public void InitializeTarget()
        {
           
            for (int i = 0; i < this.table.Rows.Count; i++)
            {
                this.TargetAttr.Add(this.table.Rows[i].ItemArray[comboBox1.SelectedIndex]);
            }

 
        }


        public void DevideSample()
        {
            double n = 0;
            switch (comboBox2.SelectedIndex)
            {
                case 0:
                    n = 0.9;
                    break;
                case 1:
                    n = 0.8;
                    break;
                case 2:
                    n = 0.7;
                    break;
                case 3:
                    n = 0.6;
                    break;
                case 4:
                    n = 0.5;
                    break;
                case 5:
                    n = 0.4;
                    break;
                case 6:
                    n = 0.3;
                    break;
                case 7:
                    n = 0.2;
                    break;
                case 8:
                    n = 0.1;
                    break;
                default:
                    n=0.9;
                    break;
            }
            //printing teaching sample
            int num = Convert.ToInt32(this.table.Rows.Count * n);

            this.teachingTable = this.table.Clone();
            for (int i = 0; i < num; i++)
            {
                this.teachingTable.ImportRow(this.table.Rows[i]);
            }
            dataGridView1.DataSource = this.teachingTable;
            //initializing training sample
            foreach (DataRow row in this.teachingTable.Rows)
            {
                ArrayList itemsArray = new ArrayList();
                {
                    foreach (string value in row.ItemArray)
                    {
                        itemsArray.Add(value);
                    }
                    this.trainingSamples.Add(new Sample(itemsArray, comboBox1.SelectedIndex));
                }
            }

            //printing test sample
            this.testTable = this.table.Clone();
            for (int i = 0; i < this.table.Rows.Count - num; i++)
            {
                this.testTable.ImportRow(this.table.Rows[this.table.Rows.Count-1-i]);
            }

            foreach (DataRow row in this.testTable.Rows)
            {
                answers.Add(row[comboBox1.SelectedIndex]);
                row[comboBox1.SelectedIndex] = "?";
            }

                dataGridView2.DataSource = this.testTable;

            //initializing test sample
                foreach (DataRow row in this.testTable.Rows)
                {
                    ArrayList itemsArray = new ArrayList();
                    foreach (string value in row.ItemArray)
                    {
                            itemsArray.Add(value);
                    }
                    this.testSamples.Add(new Sample(itemsArray, comboBox1.SelectedIndex));
                }

        }

        public ArrayList GetClosestNeighbours(int k, int j)
        {
            ArrayList tempp = new ArrayList();
            ArrayList closestNeighbours = new ArrayList();
           
            double eDistance = 100;
            {
                ArrayList TempSample = this.trainingSamples.Clone() as ArrayList;
                for (int i = k; i > 0; i--)
                {
                    ArrayList neighbour = new ArrayList();
                    foreach (Sample trainingSample in TempSample)
                    {
                        double temp = (this.testSamples[j]as Sample).CountDistance(trainingSample);
                        if (temp < eDistance)
                        {
                            
                            eDistance = temp;
                            ArrayList distance = new ArrayList();
                            distance.Add(trainingSample.GetClass());
                            distance.Add(eDistance);
                            neighbour.Clear();
                            neighbour.Add(distance);
                          
                        }
                    }
                    tempp.Add((neighbour[0]as ArrayList)[0]);
                    closestNeighbours.Add(neighbour);
                    TempSample.Remove((neighbour[0] as ArrayList)[0]);
                    eDistance = 100;
                }
            }
            return tempp;
        }



        public string GetClass(ArrayList closestNeighbours)//:)
        {
            ArrayList goal = new ArrayList();
            ArrayList targetValues = new ArrayList();
            foreach (string value in this.TargetAttr)
            {
                if (!targetValues.Contains(value))
                    targetValues.Add(value);
            }

            foreach (string value in targetValues)
            {
                ArrayList temp1 = new ArrayList();
                goal.Add(temp1);
            }

            foreach (string class_ in closestNeighbours)
            {
                for (int i = 0; i < targetValues.Count; i++)
                {
                    if (class_.Equals(targetValues[i]))
                        (goal[i] as ArrayList).Add(1);
                }
            }

            int a =0;
            for (int i = 0; i < goal.Count; i++)
            {
                if ((goal[i] as ArrayList).Count > a)
                    a = i;
            }
            return targetValues[a].ToString();
        }

        public void Randomize()
        {
            Random rnd = new Random();
            int count = this.table.Rows.Count;

            for (int i = 0; i < 4*count; i++)
            {
                DataRow row = table.NewRow();
                int rownum = rnd.Next(0, count - 1);

                row.ItemArray = table.Rows[rownum].ItemArray;
                table.Rows.RemoveAt(rownum);
                table.Rows.Add(row);
            }
        }



        public void Classify()
        {
            this.classifiedTable = this.testTable.Clone();
            for (int i = 0; i < this.testTable.Rows.Count; i++)
            {
                this.classifiedTable.ImportRow(this.testTable.Rows[i]);
            }

            for (int i = 0; i < this.testSamples.Count; i++)
            {
                ArrayList closestNeighbours = GetClosestNeighbours(Convert.ToInt32(textBox1.Text), i);
                this.classifiedTable.Rows[i][this.classifiedTable.Rows[0].ItemArray.Length - 1] = GetClass(closestNeighbours);
            }

            
        }
        public void GetAccuracy()
        {
            int match = 0;
            for (int i = 0; i < this.classifiedTable.Rows.Count; i++)
            {
                string w = this.classifiedTable.Rows[i][this.classifiedTable.Rows[0].ItemArray.Length - 1].ToString();
                if (this.classifiedTable.Rows[i][this.classifiedTable.Rows[0].ItemArray.Length - 1].Equals(answers[i]))
                    match++;
            }
            double accuracy = Convert.ToDouble(match) / Convert.ToDouble(this.classifiedTable.Rows.Count);

            label1.Text = "Точность классификатора = " + (accuracy * 100).ToString() + "%";
            label1.Visible = true;
        }

        
        public void ShowSamples()
        {
            dataGridView1.DataSource = this.teachingTable;
            dataGridView2.DataSource = this.testTable; 
        }


        private void button1_Click(object sender, EventArgs e)
        {
            OpenSample();
            InitializeTarget();
            dataGridView1.DataSource = this.table;

            button4.Enabled = true;
            groupBox1.Enabled = true;
            groupBox2.Enabled = true;
            button3.Enabled = true;
        }
        private void button2_Click(object sender, EventArgs e)
        {
            Classify();
            label1.Visible = true;
            dataGridView2.DataSource = this.classifiedTable;
            GetAccuracy();
        }

        private void button3_Click(object sender, EventArgs e)
        {
            this.testSamples.Clear();
            this.trainingSamples.Clear();
            this.answers.Clear();

            this.teachingTable.Clear();
            this.teachingTable.Rows.Clear();
            this.teachingTable.Columns.Clear();
            this.testTable.Clear();
           this.testTable.Rows.Clear();
            this.testTable.Columns.Clear();
            DevideSample();
            ShowSamples();
            button2.Enabled = true;
            groupBox3.Enabled = true;

            this.classifiedTable.Clear();
            this.classifiedTable.Rows.Clear();
            this.classifiedTable.Columns.Clear();
        }


        private void button4_Click(object sender, EventArgs e)
        {
            Randomize();
            dataGridView1.DataSource = this.table;
        }

     
       

        
     
      
    }
}