Linear-Regression

The LinearRegreesion.py is the module that I wrote in python. It uses Numpy and matplotlib. Refer try_linear_regression.py to see how you can use it.

To import the LinearRegression class Use:

 from LinearRegression import LinearRegression

Now create the instance of the class as:

 model = LinearRegression()

Using this you can use all the methods of this class:

1] Use model.fit(x, y, alpha, iteations, return_params):

  alpha--learning rate (default=0.01)
  
  iterations--iterations to train the model(default=1500)
  
  return_params--returns the trained parameters(default=False)...set return_params = True if you want the trained parameters

2] Use model.predict(X):

This method returns prediction for a give set of values of X

    X--the value of X for which you want to predict the values of y

3] Use model.plotdata(X, y,title="Title", X_label="X-axis", y_label="Y-axis"):

This method plots your dataset

  X--the x value of your data
  y--the y value of  your data
  title--the title of your plot (default=Title)
  X_label--the label of your X-axis(default=X-axis)
  y_label--The label of your Y-axis(default=Y-axis)

This is as example of how your plot might look like:

4] Use model.plotregressor(X, y,title="Title", X_label="X-axis", y_label="Y-axis")

This method plots your regressor model

  X--the x value of your data
  y--the y value of your data
  X_label--the label of your X-axis(default=X-axis)
  y_label--the label of your y-axis(default=y-axis)

This is an example of how your plot might look like:

5] Use model.plotcost()

This method plots cost function

This is an example of how your plot might look like:

Note:

1. You need to have numpy installed
2. You need to have matplotlib installed
3. This works only for the dataset where X has only one feature ---> i.e. this is simple linear regression. So if your dataset has more than one feature for X, then all the above methods will not work and will throw dimentionality error