Basic of ‘Recommender System’ ML :
A recommender system predicts users interest and recommends product that the user may be interested in .
Gathers data from :
- Explicit ratings that users provides.
- Implicit search engine queries and purchased histories.
- Other source of knowledge about the users/ item.
Recommender system is an information-filtering technique that provides users with recommendations for which they might be interested in.
Recommender system act as a solution for your day to day choices.
- Which websites will you find interesting ?
- Which degree and university are best for your future ?
- -Which is the investment for supporting the education of your children ?
Paradigms of Recommender System :
- Reduce the information overload by estimating relevance.
- Personalised recommendation : Recommends the most relevant item based on user profile and contextual parameters.
- Collaboration filtering recommendation : Recommends what’s popular among your peers.
- Content based recommendation : Displays products similar to the products you have like before.
- Knowledge – based recommender system : Recommends the products based on user’s requirement.
Collaboration Filtering :
Collaboration filtering system : Collaboration filtering system makes recommendations based on historic preferences of the users.
User – based Nearest Neighbour : User – based Nearest Neighbour recommends items by finding user to active user.
Measuring User Similarity : Pearson Correlations
- Pearson Correlations measure of how strong a relationship is between two variables.
- Degree of linearity can be determined by using Pearson Correlation.
- It determines linear component of association between two continuous variables.
- It has no assumptions about linearity.
- Pearson’s correlation is not used to determine the strength.
To read “All About Supervised Learning” : https://seeve.medium.com/all-about-supervised-learning-dbe9fe9fb676