Download the csv file and Use only python and basic libraries such as numpy,pandas,matplotlib to implement the following.
1. Split the data into train(80%) and test(20%) data randomly.
2. Use "logistic regression"(Any traditional ML method is fine but do not use Neural networks) for prediction.
3. Train : Print intermediate training loss and correctness.
- (Optionally plot training loss and correctness.)
4. Report accuracy on test data.
5. Save the code on github repo you shared with appropriate name: jupyter notebook with all outputs("pclub_logistic.ipynb",highly recommended) OR python file("pclub_logistic.py") + Output file("pclub_output.txt") containing:
a. Intermediate results during training for each epoch:
- e.g. Epoch 2 : (Loss: 0.342,Correctness: 0.69)
b. Accuracy on test dataset in format: Correct/Total (%correct)
- e.g. 363/400 (90.75%)
Optionally include plots in "pclub_plots" folder if you have any.
Running python file should generate the same output you shared ( accuracy + optionally plots.)
Accuracy Will not be used judged. Incomplete implementation is fine as long as it shows your effort. Use any Output format but ensure they are readable.