Recall rate, accuracy rate, F1 analysis 4 days before a contest, purchase by shopping cart, favorites, Shopping cart & favourites

4 days prior to purchase via shopping cart

precision:0.0153313777615
recall:0.177594903106
f1:0.0282260616417

4 days prior to purchase via favorite folder

precision:0.00326368668608
recall:0.0270772498009
f1:0.00582524271845

4 days prior to purchase via shopping cart & favourites

precision:0.0166666666667
recall:0.00955667640032
f1:0.0121477982116

Two days ago, purchased via shopping cart

precision:0.0256597238047
recall:0.149455800372
f1:0.0437995954567

Two days ago, purchased through favorites

precision:0.00478835472132
recall:0.0199097425007
f1:0.00772002058672

2 days prior to purchase via shopping cart & favourites

precision:0.0256166982922
recall:0.00716750730024
f1:0.0112009956441

Purchase by shopping cart the day before

precision:0.0310559006211
recall:0.0876028670029
f1:0.0458556242618

The day before, buy through favorites

precision:0.00696774193548
recall:0.0143350146005
f1:0.0093774420422

The day before, purchase via shopping cart & favourites

precision:0.0308370044053
recall:0.00371648526679
f1:0.00663349917081

Can be seen, only one day data for the shopping cart behavior is the best, the collection folder in the short term is a pit drop, even take intersection is also a pit drop goods.

So, keep a day of data, shopping cart data

Recall rate, accuracy rate, F1 analysis 4 days before a contest, purchase by shopping cart, favorites, Shopping cart & favourites