Most recently, Microsoft has introduced the ML.NET Model Builder tool to support machine learning predictive system processing, but it has only been introduced in the general model, so it is still difficult to apply into actual house price predictions. Therefore, we applied and improved this model to develop a new model ML.RealEstate that focuses on the private housing price prediction segment. The model will show the most common characteristics for private houses, the model will help people estimate the actual price of the house to avoid financial loss when inflated.

This Research from KMOU (Korea Maritime & Ocean University) – Data Science Lab – Room 407.

Authors: Duy Thanh Tran, Prof. Jun-Ho Huh

Any question, please free to contact me: thanhtd@uel.edu.vn

My full name: TRAN DUY THANH

Blog study coding: https://duythanhcse.wordpress.com/

Group support: https://www.facebook.com/groups/communityuni/

ML.RealEstate – How to use?

Install nuget package

Install-Package ML.RealEstate -ProjectName YourProject

The classes of the ML.RealEstate:

Download and Copy Dataset folder “RealEstateDataset”(https://github.com/thanhtd32/ML.RealEstate/tree/main/RealEstateDataset) into your execution directory. We will continue update the dataset day by day, now there are about 2.353 private house transactions, in the future the transactions have more than 10.000 transactions , just like this figure as below:

All Soure code how to use ML.RealEstate is stored https://github.com/thanhtd32/ML.RealEstate/tree/main/ML.RealEstateDemo, you can download all from https://github.com/thanhtd32/ML.RealEstate

using ML.RealEstate.Data;
using ML.RealEstate.Predict;
using System.IO;
namespace ML.RealEstateDemo
{
    public partial class frmMain : Form
    {
        BrokerRealEstate broker = new BrokerRealEstate();
        string folder = "Models";
        public frmMain()
        {
            InitializeComponent();
        }

        private void frmMain_Load(object sender, EventArgs e)
        {
            RealEstateDatabase.LoadAllDataset();
            ShowDataInUI();
            LoadModelIntoCombo();
        }
        private void LoadModelIntoCombo()
        {
            cboModel.Items.Clear();
            
            if(Directory.Exists(folder)==false)
            {
                return;
            }
            string []files=Directory.GetFiles(folder);
            foreach (string file in files)
            {
                FileInfo fi=new FileInfo(file);
                cboModel.Items.Add(fi.Name);
            }
        }

        private void ShowDataInUI()
        {
            cboHouseType.DataSource = RealEstateDatabase.GetHouseTypes();
            cboHouseType.ValueMember = "Id";
            cboHouseType.DisplayMember = "TypeOfHouse";

            cboCity.DataSource = RealEstateDatabase.GetCities();
            cboCity.ValueMember = "Id";
            cboCity.DisplayMember = "CityName";
        }
        //Step 1. Import Data and create Train - Test Set
        private void btnImportData_Click(object sender, EventArgs e)
        {
            double ratio = double.Parse(txtRatio.Text);
            int seed = int.Parse(txtSeed.Text);
            bool ret = broker.ImportDataset(RealEstateDatabase.HouseDataList!, ratio, seed);
            if (ret)
                lblStatusImportData.Text = "Import and make train - test dataset successfully";
            else
                lblStatusImportData.Text = "Import and make train - test dataset failed";
            lblStatusBuildModel.Text = "";
            lblStatusEvaluate.Text = "";
            lblStatusSaveModel.Text = "";
            lblStatusLoadModel.Text = "";
        }
        //Step 2. Build Model
        private void btnBuildModel_Click(object sender, EventArgs e)
        {
            string[] features ={ "HouseTypeId", "WardId", "DistrictId", "CityId",
                    "Area", "FrontiSpiece","Entrance","Floor","BedRoom","ToiletRoom" };
            int iterator = 100;
            bool ret = broker.BuildModel(features, iterator);
            if (ret)
                lblStatusBuildModel.Text = "Build Model successfully";
            else
                lblStatusBuildModel.Text = "Build Model failed";
        }
        //Step 3. Evaluate
        private void btnEvaluate_Click(object sender, EventArgs e)
        {
            Metric metric = broker.Evaluate();
            txtRSquared.Text = metric.RSquared.ToString();
            txtMSE.Text = metric.MSE.ToString();
            txtRMSE.Text = metric.RMSE.ToString();
            txtMAE.Text = metric.MAE.ToString();
            txtLossFunction.Text = metric.LossFunction.ToString();
        }
        //Step 4. Save Model
        private void btnSaveModel_Click(object sender, EventArgs e)
        {
            if(Directory.Exists(folder)==false)
            {
                Directory.CreateDirectory(folder);  
            }
            string path = folder + "\\ML.RealEstateModel-"+DateTime.Now.ToString("ddMMyyyy-hhmmss")+".zip";
            bool ret = broker.SaveModel(path);
            if (ret)
                lblStatusSaveModel.Text = "Save Model successfully";
            else
                lblStatusSaveModel.Text = "Save Model failed";
            LoadModelIntoCombo();
        }
        //Step 5. Load Model
        private void btnLoadModel_Click(object sender, EventArgs e)
        {
            if (cboModel.SelectedIndex == -1)
                return;
            string modelName = folder + "\\" + cboModel.Text;
            bool ret = broker.LoadModel(modelName);
            if (ret)
                lblStatusLoadModel.Text = "Load Model successfully";
            else
                lblStatusLoadModel.Text = "Load Model failed";
        }
        //6.Predict
        private void btnPredict_Click(object sender, EventArgs e)
        {
            if (cboDistrict.SelectedItem == null ||
                cboCity.SelectedItem == null)
                return;
            House inputData = new House();
            inputData.HouseTypeId = (cboHouseType.SelectedItem as HouseType)!.Id;
            inputData.WardId = (cboWard.SelectedItem as Ward)!.Id;
            inputData.DistrictId = (cboDistrict.SelectedItem as District)!.Id;
            inputData.CityId = (cboCity.SelectedItem as City)!.Id;
            inputData.Area = float.Parse(txtArea.Text);
            inputData.FrontiSpiece = float.Parse(txtFrontiSpiece.Text);
            inputData.Entrance = float.Parse(txtEntrance.Text);
            inputData.Floor = float.Parse(txtFloor.Text);
            inputData.BedRoom = float.Parse(txtBedRoom.Text);
            inputData.ToiletRoom = float.Parse(txtToiletRoom.Text);
            Prediction result = broker.Predict(inputData);
            txtPrice.Text = result.Price.ToString();

           /* lblStatusBuildModel.Text = "";
            lblStatusEvaluate.Text = "";
            lblStatusSaveModel.Text = "";
            lblStatusLoadModel.Text = "";
            lblStatusImportData.Text = ""; */
        }

        private void cboCity_SelectedIndexChanged(object sender, EventArgs e)
        {
            if (cboCity.SelectedIndex == -1)
                return;
            City? city = cboCity.SelectedItem as City;
            cboDistrict.DataSource = RealEstateDatabase.GetDistricts(city!.Id);
            cboDistrict.DisplayMember = "DistrictName";
            cboDistrict.ValueMember = "Id";
        }

        private void cboDistrict_SelectedIndexChanged(object sender, EventArgs e)
        {
            if (cboDistrict.SelectedIndex == -1)
                return;
            District? district = cboDistrict.SelectedItem as District;
            cboWard.DataSource = RealEstateDatabase.GetWards(district!.Id);
            cboWard.DisplayMember = "WardName";
            cboWard.ValueMember = "Id";
        }
    }
}