我试图按类别分类一系列文本示例新闻 . 我有大量数据集的新闻文本与数据库中的类别 . 机器应该训练并决定新闻类别 .
public static string[] Tokenize(string text)
{
StringBuilder sb = new StringBuilder(text);
char[] invalid = "!-;':'\",.?\n\r\t".ToCharArray();
for (int i = 0; i < invalid.Length; i++)
sb.Replace(invalid[i], ' ');
return sb.ToString().Split(new[] { ' ' }, System.StringSplitOptions.RemoveEmptyEntries);
}
private void Form1_Load(object sender, EventArgs e)
{
string strDSN = "Provider=Microsoft.ACE.OLEDB.12.0;Data Source = c:\\users\\158820\\Documents\\Database4.accdb";
string strSQL = "SELECT * FROM NewsRepository";
// create Objects of ADOConnection and ADOCommand
OleDbConnection myConn = new OleDbConnection(strDSN);
OleDbDataAdapter myCmd = new OleDbDataAdapter(strSQL, myConn);
myConn.Open();
DataSet dtSet = new DataSet();
myCmd.Fill(dtSet, "NewsRepository");
DataTable dTable = dtSet.Tables[0];
myConn.Close();
StringBuilder sWords = new StringBuilder();
string[][] swords = new string[dTable.Rows.Count][];
int i = 0;
foreach (DataRowView dr in dTable.DefaultView)
{
swords[i] = Tokenize(dr[1].ToString());
i++;
}
Codification codebook = new Codification(dTable, new string[] { "NewsTitle", "Category" });
DataTable symbols = codebook.Apply(dTable);
int[][] inputs = symbols.ToJagged<int>(new string[] { "NewsTitle" });
int[] outputs = symbols.ToArray<int>("Category");
bagOfWords(inputs, outputs);
}
private static void bagOfWords(int[][] inputs, int[] outputs)
{
var bow = new BagOfWords<int>();
var quantizer = bow.Learn(inputs);
string filenamebow = Path.Combine(Application.StartupPath, "News_BOW.accord");
Serializer.Save(obj: bow, path: filenamebow);
double[][] histograms = quantizer.Transform(inputs);
// One way to perform sequence classification with an SVM is to use
// a kernel defined over sequences, such as DynamicTimeWarping.
// Create the multi-class learning algorithm as one-vs-one with DTW:
var teacher = new MulticlassSupportVectorLearning<ChiSquare, double[]>()
{
Learner = (p) => new SequentialMinimalOptimization<ChiSquare, double[]>()
{
// Complexity = 100 // Create a hard SVM
}
};
// Learn a multi-label SVM using the teacher
var svm = teacher.Learn(histograms, outputs);
// Get the predictions for the inputs
int[] predicted = svm.Decide(histograms);
// Create a confusion matrix to check the quality of the predictions:
var cm = new GeneralConfusionMatrix(predicted: predicted, expected: outputs);
// Check the accuracy measure:
double accuracy = cm.Accuracy;
string filename = Path.Combine(Application.StartupPath, "News_SVM.accord");
Serializer.Save(obj: svm, path: filename);
}
我对如何训练accord.net对象感到困惑 . 我能够序列化经过训练的模型(在9个类别中,3600个独特新闻大约为106 MB)
如何使用该模型预测新的一组新闻文本的类别?
1 回答
使用您的模型不在训练集中的数据就像调用您的svm做出另一个决定一样简单:
由于您已经对已训练的模型进行了序列化,因此可以使用
Serializer.Load<T>
实例化svm对象,该文档记录为here .