Description: The aim is to design and develop an algorithm which is capable to classify the financial transactions of a person. Given are the data for the transactions and the user specified classes. The algorithm should be intelligent, for example, the accent marks might be missing from the comment of a transaction. The algorithm should be able to learn. For example, if it could not classify the flour and I did it manually, then next time it should be done automatically. For each item a primary and secondary classes should be proposed. Naturally, the user should be able to override the decision of the algorithm. Sometime, one transaction includes several items, a solution should be proposed for this situation.
Supervisor: Heckl István