Goal

• Learn to improve a multi-layer perceptron by learning.

• Methods of learning improvement: reducing overfitting, regularization, early stop, dropout.

Readings

• R1 - Chapter 21 Neural Networks,  Sections 21.8~21.14, Machine Learning with Python Cookbook 2nd,  Free O'Reilly Learning by FJU ID

Instructions

• Read R1 and practice all the Python codes.

• Exercise: Use at least two datasets of Scikit-Learn real-world dataset in your exercise. Modify the Python codes to perform neural networks on new data.  Analyze the results of output, explain the results, and improve the results. Compare the results between P8 and P9.

Report: a separate and single web page that you

• • Write the goal of this assignment

  • Paste the Python codes you write. Clearly explain modified code segments written by you.

  • Upload code output results into your report web page

  • Explain the results of the output results. Compare the results between P8 and P9.

  • Give references to this assignment (teacher's readings, your search and readings)

Submit the web address of your web page to Microsoft Teams.