Use a transaction dataset to build a classification algorithm predicting client purchases in the next 3 months.

Q1. From the data (transaction date, client id, color of item, amount), what features would you create to build the model? Q2. You train a model and you get a “too good to be true” accuracy, what could have happened? Q3. You build an SVC model and the accuracy is not very good. What could you do to improve your model?