Hey Matt, you have included total number of times a customer has visited the store in the training dataset. There is a 2 year tenure customer, and there is a 3 month tenure customer, both have similar purchasing patterns. But the model might not capture this, because the frequency is being taken in absolute terms. How do you address this?
Even if the average number of frequencies is considered for training data, customer who purchases irregularly might not make a purchase in the next 90 days, how do you capture this buying patterns? Irregular purchase patterns.
Hey Matt, In reg.predict_model, you have included the dataframe that contains the target variable "sales_90_value", isn't it obvious that the model will be easily influenced by this value and predict almost nearest values as the actual values? shouldn't the target variable be excluded from the data when passing it for prediction? Can you please clarify here why you have included it. Thank You😄
Hey Matt, you have included total number of times a customer has visited the store in the training dataset. There is a 2 year tenure customer, and there is a 3 month tenure customer, both have similar purchasing patterns. But the model might not capture this, because the frequency is being taken in absolute terms. How do you address this?
Even if the average number of frequencies is considered for training data, customer who purchases irregularly might not make a purchase in the next 90 days, how do you capture this buying patterns? Irregular purchase patterns.
Hey Matt,
In reg.predict_model, you have included the dataframe that contains the target variable "sales_90_value", isn't it obvious that the model will be easily influenced by this value and predict almost nearest values as the actual values? shouldn't the target variable be excluded from the data when passing it for prediction? Can you please clarify here why you have included it.
Thank You😄
It’s not included in the predictors. So the xgboost model won’t use it. Only as a regression label.