lasso_scratch.py

df=pd.read_excel(r'C://Users//Prashant//Desktop//airfoil_new.xlsx')
columns = ['Frequency','Angle','Chord length','Free-stream velocity','Suction side displacement thickness','Scaled sound pressure level']
df.columns=columns


class LassoRegression():
    
    def __init__( self, lr, iterations, l1_penality ):
        
        self.lr = lr 
        self.iterations = iterations 
        self.l1_penality = l1_penality 
          
    # Function for model training 
              
    def fit( self, X, Y ):
         
        self.m, self.n = X.shape  #rows,columns
          
        # weight initialization 
        self.W = np.zeros( self.n ) 
        self.b = 0
        self.X = X 
        self.Y = Y 
          
        # optimization learning using the helper function 
                  
        for i in range( self.iterations ):
            self.update_weights() 
              
        return self
      
    # Helper function to update weights in gradient descent 
      
    def update_weights( self ):
        
        Y_pred = self.predict(self.X) 
        dW = np.zeros( self.n ) # zero array of the size of columns of training seyt
        
        for j in range( self.n ):
            
            if self.W[j] > 0:
                #calculating the differential of the expression for weight values>0
                dW[j] = (-(2*(self.X.iloc[:, j]).dot(self.Y - Y_pred))+self.l1_penality)/self.m  
          
            else :
                #calculating the differential of the expression for weight values<0
                dW[j] = ( - ( 2 * ( self.X.iloc[:,j] ).dot( self.Y - Y_pred ) ) - self.l1_penality ) / self.m  
  
       
        db = - 2 * np.sum( self.Y - Y_pred )/self.m  #taking the differential of the error function  
          
        # update weights 
      
        self.W = self.W - self.lr * dW 
      
        self.b = self.b - self.lr * db 
          
        return self
      
    # prediction function
      
    def predict( self, X ):
        return X.dot( self.W ) + self.b 
      
  
if __name__ == '__main__':
    
    model = LassoRegression( iterations = 1000, lr = 0.01, l1_penality = 50 )
    model.fit( X_train, y_train ) 
    Y_pred = model.predict( X_test ) 
    print( "Predicted values ", np.round( Y_pred, 2 ))