SlopeOne is an Item-Based recommendation algorithm proposed by & nbsp; DanielLemire & nbsp; professor at & nbsp; 2005 & nbsp. & Nbsp; The SlopeOne algorithm tries to meet the following five goals at the same time: easy implementation and maintenance: common engineers can easily explain all the aggregated data, and the algorithm is easy to implement and test.
What is the Slope One algorithm? Daniel Lemire? Professor? 2005? .? The Slope One algorithm tries to meet these five goals at the same time:
- Easy to implement and maintain: common engineers can easily interpret all aggregated data, and the algorithms are easy to implement and test.
- Which can be updated during running: a new score item should have immediate impact on the prediction result.
- Efficient query response: fast query execution may require more space occupation.
- Less requirements for first-time visitors: effective recommendations should also be available to users with few scoring items.
- Reasonable accuracy: compared with the most accurate method, this method should be competitive. the slight increase in accuracy cannot be at the cost of simplicity and scalability.
Use this figure to briefly describe the Slope One algorithm.
- User A scores Item I as 1 and Item J as 1.5.
- Uesr B scores Item I as 2.
- Question: How much does User B score Item J?
- When using the Slope One algorithm, the answer is: 2.5, 2 + (1.5-1) = 2.5.
The Slope One algorithm is so simple. for detailed test analysis, see "Slope One Predictors for Online Rating-Based Collaborative Filtering ".
