Assignment 4
— Neural Network System for Stock Advisory You are a Machine Learning expert in a consultancy company, providing ML service to various industries and clients.
A financial institution commissioned your company a project to analyze their confidential data on the performance of all the stocks they have been watching or are in their portfolios. In order to kept their data confidential, the key 10 features which reflects the “environment” of the stock market (they are valuable properties of the company) are anonymized and named as “Fl”, “F2″…”F10”. From their past data and experience, they provide you with 3 classes of outcomes – “BUY”, “HOLD” or “SELL”.
Your assignment is to use neural network to build a system which can be used by the financial institute to assist their investment experts.
Two datasets in separate files consist of 9,000 instances for training and 1,000 instances for testing. The following is the data dictionary:
Fl, F2 …. F10 Feature sets provided by the customer class 0 = “SELL”; 1 = “HOLD”; 2 = “BUY”
You should include the following in your report:
1. Brief analysis of the dataset.
2. Design and build your neural network, how you train and test it. 3. Any observation you may have in your project.
Your report should be a brief write up and your Colab notebook with executable codes as appendix.
You may choose to include your report within the Colab notebook as collection of text cells, although that may be more difficult to format.
Datasets provided: stock_market_training.csv, stock_market_testing.csv
