How to increase efficiency of speech recognition using system.speech class in c#

See more:

C#

C#4.0

speech

recognition

I'm doing speech recognition using c# with system.speech library but the quality is not so good i have to recognize about 4000 english words.this is the code

private void Form1_Load(object sender, EventArgs e)
                {
              Choices commands = new Choices();
              GrammarBuilder gb = new GrammarBuilder();
 
 
 
   
           
   
            commands.Add(new string[] { "red", "blue", "green" ,"meantime","transmit","harvest","consecutive","coordinate","spy","slot","riot","nutrient","citizenship","severely","sovereignty","ridge","brave","lighting","specify","contributor","frustrate","articulate","importantly","transit","dense","seminar","electronics","sunny","shorts","swell","accusation","soften","straighten","terribly","cue","bride","biography","hazard","compelling","seldom","tile","economically","honestly","troubled","twentieth","balanced","foreigner","convenience","delight","weave","timber","till","accurately","plea","bulb","flying","sustainable","devil","bolt","cargo","spine","seller","skilled","managing","marine","dock","organized","fog","diplomat","boring","sometime","summary","missionary","epidemic","fatal","trim","warehouse","accelerate","butterfly","bronze","drown","inherent","nationwide","spit","kneel","vacuum","selected","dictate"});
            
 
 
            
            gb.Append(commands);
          
            Grammar grammar = new Grammar(gb);
 
            recEngine.LoadGrammarAsync(grammar);
            recEngine.SetInputToDefaultAudioDevice();
 
            recEngine.SpeechRecognized+=recEngine_SpeechRecognized;
 
            recEngine.RecognizeAsync(RecognizeMode.Multiple);
}
 
 
void recEngine_SpeechRecognized(object sender,SpeechRecognizedEventArgs e)
      {
 
          switch(e.Result.Text){
 
              
                  textBox1.Text += e.Result.Text;
                 

 
          }<pre>
 


Is there any way to improve efficiency through training ...do windows default speech recognition  accessed through control panel/speech recognition  traing helps to improve accuracy what are other modes for training