DataMining-Homework

This project is focused on the extraction of association rules from a dataset of transactions using two Market Basket Analysis algorithms: Apriori and FP-Growth.

Technologies used

• Pandas: data manipulation
• Mlxtend: Apriori, FP-Growht and association rules function
• Matplotlib: data visualization

I developed three Python modules to make more easy the implementation:

• "dataCleaning"
• "preAnalisys"
• "data_rappresentation".

Data Preprocessing

To have good results, the dataset needs to was cleaned by:

• Nan values
• Unuseless values
• Product with low support
• Transaction with only one product

After filtering, the dataset was trasformed into binary transaction matrix

• 1 means the product is present in the transaction
• 0 means it

To do that, I have developed the dataCleaning and preAnalisys libraries that includes all these operations

Data visualization

To make more easy and speedy the data visualization I have developed data_rappresentation. That was created using Matplotlib

Marlet Basket Analisys

After all the preprocessing operations I can finally apply

• Apriori
• FP-Growht After generating frequent itemsets, the association rules can be extracted