//Hàm dùng để tại ra các khoảng giá trị đã rời rạc
public static void TaoBayesObject(String colName, int khoang, DataTable dt)
{
BayesObjectTableAdapter BayesObjectTA = new BayesObjectTableAdapter();
//DataSetTableAdapter dataSetTA = new DataSetTableAdapter();
//DataTable dt = dataSetTA.GetData();
decimal minValue = Convert.ToDecimal(dt.Compute("min(" + colName + ")", string.Empty));
decimal maxValue = Convert.ToDecimal(dt.Compute("max(" + colName + ")", string.Empty));
decimal giaTriMoi = Math.Round((minValue + maxValue) / khoang, 3);
decimal giaTriTrungBinhKhoang = giaTriMoi;
for (int i = 1; i <= khoang; i++)
{
String khoangRoiRac = "";
if (i == 1)
{
khoangRoiRac = "[0," + giaTriMoi.ToString("0.###") + ")";
}
else if (i == khoang)
{
khoangRoiRac = "[" + giaTriMoi.ToString("0.###") + ",+)";
}
else
{
khoangRoiRac = "[" + (giaTriMoi - giaTriTrungBinhKhoang).ToString("0.###") + ","
+ giaTriMoi.ToString("0.###") + ")";
}
BayesObjectTA.Insert(colName, khoangRoiRac, 0, "True");
BayesObjectTA.Insert(colName, khoangRoiRac, 0, "False");
if (i != khoang - 1)
{
giaTriMoi = giaTriMoi + giaTriTrungBinhKhoang;
}
}
}
//Hàm dùng để lấy một tập gồm các khoảng rời rạc của một cột
public static List <double> GetListIntervalValue(String colName)
{
List <double> listIntervalValues = new List <double>();
if (colName != "GioiTinh")
{
BayesObjectTableAdapter bayesObjectTableAdapter = new BayesObjectTableAdapter();
DataRow[] dtIntervalValue = bayesObjectTableAdapter.GetData().Select("TenThuocTinh='" + colName + "'");
foreach (DataRow dtRow in dtIntervalValue)
{
String[] listString = dtRow[2].ToString().Split(',');
String str = listString[0].Substring(1, listString[0].Length - 1);
double intervalValue = Convert.ToDouble(str);
if (!listIntervalValues.Contains(intervalValue))
{
listIntervalValues.Add(intervalValue);
}
//Dành cho trường hợp thuộc tính chỉ có 1 khoảng
if (dtIntervalValue.Count() == 2)
{
String secondStr = listString[1].Substring(0, listString[1].Length - 1);
double secondIntervalValue = Convert.ToDouble(secondStr);
if (!listIntervalValues.Contains(secondIntervalValue))
{
listIntervalValues.Add(secondIntervalValue);
}
}
}
}
return(listIntervalValues);
}
//Hàm dùng để huấn luyện dữ liệu dành cho thuật toán Naive Bayes
public static void HuanLuyenBayes(int row)
{
BayesObjectTableAdapter dtBayesAdapter = new BayesObjectTableAdapter();
DataSetTempTableAdapter dtSetTempTA = new DataSetTempTableAdapter();
int rowCount = row;
DataTable dataForTraining = dtSetTempTA.GetDataByNumber(rowCount);
DiabetesDataSet.BayesObjectDataTable dtBayes = dtBayesAdapter.GetData();
foreach (DataRow dtRow in dtBayes.Rows)
{
String colName = dtRow[1].ToString();
String khoangRoiRac = dtRow[2].ToString();
String tieuDuong = dtRow[4].ToString();
String iQuery = "" + colName + "='" + khoangRoiRac + "' and TieuDuong='" + tieuDuong + "'";
int soLuong = dataForTraining.Select(iQuery).Count();
if (soLuong == 0)
{
soLuong = 1;
}
dtRow[3] = soLuong;
dtBayesAdapter.Update((dtBayes));
}
}
//Hàm dùng để tính kết quả của một bảng trong bộ thử nghiệm
public static DataTable NaiveBayes(DataTable dtTestSet)
{
BayesObjectTableAdapter bayesTA = new BayesObjectTableAdapter();
DataSetTempTableAdapter dataSetTempTA = new DataSetTempTableAdapter();
int dataForTrainingRowsCount = (int)dataSetTempTA.CountRows() - dtTestSet.Rows.Count;
DataTable dtBayes = bayesTA.GetData();
DataTable dtdataSetTemp = dataSetTempTA.GetDataByNumber(dataForTrainingRowsCount);
int possiveNumber = dtdataSetTemp.Select("TieuDuong='True'").Count();
int negativeNumber = dtdataSetTemp.Select("TieuDuong='False'").Count();
int allNumber = possiveNumber + negativeNumber;
int indexLastColumn = dtTestSet.Columns.Count - 1;
foreach (DataRow dtRow in dtTestSet.Rows)
{
Decimal pColTrue = 1;
Decimal pColFalse = 1;
for (int i = 1; i < dtTestSet.Columns.Count - 2; i++)
{
String colName = dtTestSet.Columns[i].ColumnName;
String khoangRoiRac = dtRow[i].ToString();
DataRow possiveRow = dtBayes.Select("TenThuocTinh='" + colName + "' and KhoangRoiRac='" + khoangRoiRac + "' and TieuDuong='True'")[0];
DataRow negativeRow = dtBayes.Select("TenThuocTinh='" + colName + "' and KhoangRoiRac='" + khoangRoiRac + "' and TieuDuong='False'")[0];
Decimal ColPossiveNumber = Convert.ToDecimal(possiveRow[3]);
Decimal ColNegativeNumber = Convert.ToDecimal(negativeRow[3]);
pColTrue = pColTrue * ColPossiveNumber / possiveNumber;
pColFalse = pColFalse * ColNegativeNumber / negativeNumber;
}
Decimal pRowTrue = pColTrue * possiveNumber / allNumber;
Decimal pRowFalse = pColFalse * negativeNumber / allNumber;
if (pRowFalse == 0)
{
dtRow["TieuDuong"] = "False";
}
else
{
if (pRowTrue > pRowFalse)
{
dtRow["TieuDuong"] = "True";
}
else
{
dtRow["TieuDuong"] = "False";
}
}
}
return(dtTestSet);
}
private void buttonXDataDiscretization_Click(object sender, EventArgs e)
{
int illegalInputCount = 0;
BayesObjectTableAdapter bayesObjectTableAdapter = new BayesObjectTableAdapter();
DataSetTempTableAdapter dataSetTempTableAdapter = new DataSetTempTableAdapter();
DataSetTableAdapter dataSetTableAdapter = new DataSetTableAdapter();
for (int i = listIntervalValue.Count - 1; i > 0; i--)
{
double currentValue = listIntervalValue[i];
double previousValue = listIntervalValue[i - 1];
if (currentValue <= previousValue)
{
illegalInputCount++;
break;
}
}
if (illegalInputCount > 0)
{
MessageBox.Show("Giá trị của từng khoảng nhập vào không hợp lệ", "Thông báo", MessageBoxButtons.OK, MessageBoxIcon.Error);
}
else
{
bayesObjectTableAdapter.DeleteByOne(this.ColName);
DataTable dtDataSetForPreProcessing = dataSetTableAdapter.GetData();
DiabetesDataSet.DataSetTempDataTable dtDataSetTemp = dataSetTempTableAdapter.GetData();
int rowIndex = 0;
foreach (DataRow dtRow in dtDataSetForPreProcessing.Rows)
{
decimal colValue = Convert.ToDecimal(dtRow[this.ColName]);
decimal maBn = Convert.ToDecimal(dtRow["MaBn"]);
String discretizationValue = DataDiscretization(colValue, this.listIntervalValue);
dtDataSetTemp.Rows[rowIndex][this.ColName] = discretizationValue;
rowIndex++;
}
dataSetTempTableAdapter.Update(dtDataSetTemp);
CreateBayesObject(this.ColName, this.listIntervalValue);
}
this.Close();
}
private void buttonXDataDiscretization_Click(object sender, EventArgs e)
{
int illegalInputCount = 0;
BayesObjectTableAdapter bayesObjectTableAdapter = new BayesObjectTableAdapter();
DataSetTempTableAdapter dataSetTempTableAdapter = new DataSetTempTableAdapter();
DataSetTableAdapter dataSetTableAdapter = new DataSetTableAdapter();
for (int i = listIntervalValue.Count - 1; i > 0; i--)
{
double currentValue = listIntervalValue[i];
double previousValue = listIntervalValue[i - 1];
if (currentValue <= previousValue)
{
illegalInputCount++;
break;
}
}
if (illegalInputCount > 0)
MessageBox.Show("Giá trị của từng khoảng nhập vào không hợp lệ", "Thông báo", MessageBoxButtons.OK, MessageBoxIcon.Error);
else
{
bayesObjectTableAdapter.DeleteByOne(this.ColName);
DataTable dtDataSetForPreProcessing = dataSetTableAdapter.GetData();
DiabetesDataSet.DataSetTempDataTable dtDataSetTemp = dataSetTempTableAdapter.GetData();
int rowIndex = 0;
foreach (DataRow dtRow in dtDataSetForPreProcessing.Rows)
{
decimal colValue = Convert.ToDecimal(dtRow[this.ColName]);
decimal maBn = Convert.ToDecimal(dtRow["MaBn"]);
String discretizationValue = DataDiscretization(colValue, this.listIntervalValue);
dtDataSetTemp.Rows[rowIndex][this.ColName] = discretizationValue;
rowIndex++;
}
dataSetTempTableAdapter.Update(dtDataSetTemp);
CreateBayesObject(this.ColName, this.listIntervalValue);
}
this.Close();
}
//Hàm dùng để tính kết quả của một bảng trong bộ thử nghiệm
public static DataTable NaiveBayes(DataTable dtTestSet)
{
BayesObjectTableAdapter bayesTA = new BayesObjectTableAdapter();
DataSetTempTableAdapter dataSetTempTA = new DataSetTempTableAdapter();
int dataForTrainingRowsCount = (int)dataSetTempTA.CountRows() - dtTestSet.Rows.Count;
DataTable dtBayes = bayesTA.GetData();
DataTable dtdataSetTemp = dataSetTempTA.GetDataByNumber(dataForTrainingRowsCount);
int possiveNumber = dtdataSetTemp.Select("TieuDuong='True'").Count();
int negativeNumber = dtdataSetTemp.Select("TieuDuong='False'").Count();
int allNumber = possiveNumber + negativeNumber;
int indexLastColumn = dtTestSet.Columns.Count - 1;
foreach (DataRow dtRow in dtTestSet.Rows)
{
Decimal pColTrue = 1;
Decimal pColFalse = 1;
for (int i = 1; i < dtTestSet.Columns.Count - 2; i++)
{
String colName = dtTestSet.Columns[i].ColumnName;
String khoangRoiRac = dtRow[i].ToString();
DataRow possiveRow = dtBayes.Select("TenThuocTinh='" + colName + "' and KhoangRoiRac='" + khoangRoiRac + "' and TieuDuong='True'")[0];
DataRow negativeRow = dtBayes.Select("TenThuocTinh='" + colName + "' and KhoangRoiRac='" + khoangRoiRac + "' and TieuDuong='False'")[0];
Decimal ColPossiveNumber = Convert.ToDecimal(possiveRow[3]);
Decimal ColNegativeNumber = Convert.ToDecimal(negativeRow[3]);
pColTrue = pColTrue * ColPossiveNumber / possiveNumber;
pColFalse = pColFalse * ColNegativeNumber / negativeNumber;
}
Decimal pRowTrue = pColTrue * possiveNumber / allNumber;
Decimal pRowFalse = pColFalse * negativeNumber / allNumber;
if (pRowFalse == 0)
dtRow["TieuDuong"] = "False";
else
{
if (pRowTrue > pRowFalse)
dtRow["TieuDuong"] = "True";
else
dtRow["TieuDuong"] = "False";
}
}
return dtTestSet;
}
//Hàm dùng để huấn luyện dữ liệu dành cho thuật toán Naive Bayes
public static void HuanLuyenBayes(int row)
{
BayesObjectTableAdapter dtBayesAdapter = new BayesObjectTableAdapter();
DataSetTempTableAdapter dtSetTempTA = new DataSetTempTableAdapter();
int rowCount = row;
DataTable dataForTraining = dtSetTempTA.GetDataByNumber(rowCount);
DiabetesDataSet.BayesObjectDataTable dtBayes = dtBayesAdapter.GetData();
foreach (DataRow dtRow in dtBayes.Rows)
{
String colName = dtRow[1].ToString();
String khoangRoiRac = dtRow[2].ToString();
String tieuDuong = dtRow[4].ToString();
String iQuery = "" + colName + "='" + khoangRoiRac + "' and TieuDuong='" + tieuDuong + "'";
int soLuong = dataForTraining.Select(iQuery).Count();
if (soLuong == 0)
soLuong = 1;
dtRow[3] = soLuong;
dtBayesAdapter.Update((dtBayes));
}
}
private void CreateBayesObject(String colName, List <double> listInputIntervalValue)
{
BayesObjectTableAdapter bayesObjectTableAdapter = new BayesObjectTableAdapter();
for (int i = 0; i < listInputIntervalValue.Count; i++)
{
String InputIntervalValue = "";
double currentValue = listInputIntervalValue[i];
double nextValue = 0;
if (i == listInputIntervalValue.Count - 1)
{
InputIntervalValue = "[" + currentValue.ToString("0.###") + ",+)";
}
else
{
nextValue = listInputIntervalValue[i + 1];
InputIntervalValue = "[" + currentValue.ToString("0.###") + "," + nextValue.ToString("0.###") + ")";
}
bayesObjectTableAdapter.Insert(colName, InputIntervalValue, 0, TableMetaData.PositiveString);
bayesObjectTableAdapter.Insert(colName, InputIntervalValue, 0, TableMetaData.NegativeString);
}
}
private void buttonXDiscretizationDataStatistics_Click(object sender, EventArgs e)
{
if (checkedListBoxColumnName.SelectedIndex == -1)
{
MessageBox.Show("Chưa chọn thuộc tính để xem thống kê", "Thông báo", MessageBoxButtons.OK, MessageBoxIcon.Error);
}
else
{
String colName = checkedListBoxColumnName.SelectedItem.ToString();
BayesObjectTableAdapter bayesObjectTableAdapter = new BayesObjectTableAdapter();
DataSetTempTableAdapter dataSetTempTableAdapter = new DataSetTempTableAdapter();
DataTable dtBayesForStatistics = bayesObjectTableAdapter.GetData();
DataTable dtSetTempForStatistics = dataSetTempTableAdapter.GetData();
DataRow[] bayesRows = bayesObjectTableAdapter.GetData().Select("TenThuocTinh='" + colName + "'");
dtBayesForStatistics.Clear();
int iCount = 0;
foreach (DataRow bayesRow in bayesRows)
{
String tenThuocTinh = bayesRow["TenThuocTinh"].ToString();
String khoangRoiRac = bayesRow["KhoangRoiRac"].ToString();
String tieuDuong = bayesRow["TieuDuong"].ToString();
int tongSoLuong = 0;
String iQuery = "" + tenThuocTinh + "='" + khoangRoiRac + "' and TieuDuong='" + tieuDuong + "'";
tongSoLuong = dtSetTempForStatistics.Select(iQuery).Count();
DataRow newRow = dtBayesForStatistics.NewRow();
iCount++;
newRow[0] = iCount;
newRow[1] = tenThuocTinh;
newRow[2] = khoangRoiRac;
newRow[3] = tongSoLuong;
newRow[4] = tieuDuong;
dtBayesForStatistics.Rows.Add(newRow);
}
dataGridViewXDescriptiveData.DataSource = dtBayesForStatistics;
dataGridViewXDescriptiveData.Columns["ID"].Visible = true;
}
}
private void CreateBayesObject(String colName, List<double> listInputIntervalValue)
{
BayesObjectTableAdapter bayesObjectTableAdapter = new BayesObjectTableAdapter();
for (int i = 0; i < listInputIntervalValue.Count; i++)
{
String InputIntervalValue = "";
double currentValue=listInputIntervalValue[i];
double nextValue = 0;
if (i == listInputIntervalValue.Count - 1)
InputIntervalValue = "[" + currentValue.ToString("0.###") + ",+)";
else
{
nextValue = listInputIntervalValue[i + 1];
InputIntervalValue = "[" + currentValue.ToString("0.###") + "," + nextValue.ToString("0.###") + ")";
}
bayesObjectTableAdapter.Insert(colName, InputIntervalValue, 0, TableMetaData.PositiveString);
bayesObjectTableAdapter.Insert(colName, InputIntervalValue, 0, TableMetaData.NegativeString);
}
}
private void buttonXDiscretizationDataStatistics_Click(object sender, EventArgs e)
{
if (checkedListBoxColumnName.SelectedIndex == -1)
MessageBox.Show("Chưa chọn thuộc tính để xem thống kê", "Thông báo", MessageBoxButtons.OK, MessageBoxIcon.Error);
else
{
String colName = checkedListBoxColumnName.SelectedItem.ToString();
BayesObjectTableAdapter bayesObjectTableAdapter = new BayesObjectTableAdapter();
DataSetTempTableAdapter dataSetTempTableAdapter = new DataSetTempTableAdapter();
DataTable dtBayesForStatistics = bayesObjectTableAdapter.GetData();
DataTable dtSetTempForStatistics = dataSetTempTableAdapter.GetData();
DataRow[] bayesRows = bayesObjectTableAdapter.GetData().Select("TenThuocTinh='" + colName + "'");
dtBayesForStatistics.Clear();
int iCount = 0;
foreach (DataRow bayesRow in bayesRows)
{
String tenThuocTinh = bayesRow["TenThuocTinh"].ToString();
String khoangRoiRac = bayesRow["KhoangRoiRac"].ToString();
String tieuDuong = bayesRow["TieuDuong"].ToString();
int tongSoLuong = 0;
String iQuery = "" + tenThuocTinh + "='" + khoangRoiRac + "' and TieuDuong='" + tieuDuong + "'";
tongSoLuong = dtSetTempForStatistics.Select(iQuery).Count();
DataRow newRow = dtBayesForStatistics.NewRow();
iCount++;
newRow[0] = iCount;
newRow[1] = tenThuocTinh;
newRow[2] = khoangRoiRac;
newRow[3] = tongSoLuong;
newRow[4] = tieuDuong;
dtBayesForStatistics.Rows.Add(newRow);
}
dataGridViewXDescriptiveData.DataSource = dtBayesForStatistics;
dataGridViewXDescriptiveData.Columns["ID"].Visible = true;
}
}
