Referral System (Collaborative filtering, slope one)

1. Recommended algorithms in the system:

Collaborative filtering:

Based on user USER-CF

Content-based ITEM–CF

Slop One

Association rules (Apriori algorithm, beer and diapers)

2.slope One algorithm

The slope one algorithm is a personalized algorithm for predicting user ratings based on a linear algorithm of scoring differences between different items. The slope one algorithm was presented by Professor Daniel in 2005. It is divided into 2 main steps

1. Calculate the average of the difference between items, and mark the deviation of the scores between items;

2. Generate a list of recommended items with high predictive scores based on the scoring deviations between items and the user's historical score.

Instance:

3.slope One application scenario and pros and cons

7. Recommended System Evaluation indicators:

Coverage

Recall rate

F1

Referral System (Collaborative filtering, slope one)