Within the scope of this chapter, we've prepared a Python module (simple and unprepared to use in production) that implemented three different models for linear regression. Even though linear regression is often inappropriate for time-series analysis, this method is well-known and easy to understand, so it serves as the following:

• Simple: Runs the linear regression expression y=Wx + b
• Estimator: Uses TensorFlow's built-in estimator to help automate training, testing, and predicting
• Polynomial: Uses TensorFlow (this time without an Estimator) to run training and predict a polynomial linear regression expression

We can find it on the GitHub repository at https://github.com/cleveranjos/SSE-LinearRegression and download it as a ZIP file to your computer using the green Clone or download button.

Unzip the package at any location on the computer and open up the Command Prompt, where you unzip it and run the extension, as follows:

c:PythonPython36python.exe ExtensionService_LinearRegression.py

The following is what we see: