Recommender systems were taught this week where we discovered that these are software agents that obtain interests of consumers and recommend different products based on those interests. This can be done in three different ways for instance: through collaborative filtering, content-based approach and hybrid approach.
Collaborative filtering derives interest based of prior consumer behaviors or consumers with similar traits and interests to give recommendations to people. For example:
Give an assumption of a movie is given 3 stars or more it means it is liked.
The recommendations in this table are explained below are.
I would recommend romance forever to Bob because Alice gave it 5 stars that means she like liked it .
I would recommend cute puppies of love to Alice because Bob gave it 4 stars that means he liked it.
I recommend swords and karate to Dave because carol gave it 5 stars which means she liked it .
I would not recommend cute puppes of love to Dave because carol gave it 1 star which means she did not like it.
I would not reccommend Romance forever because Dave gave it one tar which means he did not like it.
Disadvantages of this method
is sparsity,some products are not rated or have few ratings, there is need for many people before there is a recommendations.
Content based filtering
The second method is content based filtering this is based on the contents consumed by users rather than their opinions. For example:
|
Blogs |
Elise |
|
Linux |
11 |
|
OpenSource |
- |
|
Cloud Computing |
9 |
|
Java Technology
|
- |
|
Agile |
1 |
linux,open source and cloud computing contain the same content elsie is interested in . considering the fact that she has read all the linux and cloud computing have been read more than five times.I would recommend Elsie to read open source because it contains contents she may be interested considering the fact she was interested in linux and cloud computing.
Example 2:
Product recommendations based on features: If a user is looking for a particular product, such as a camera or laptop, a content-based filtering system would recommend other products with similar features, such as price, brand, or technical specifications.
Disadvantage of this method
It diffcult to exploit quality judgments of others.
It requires content that can encoded as meaningful features .
Advantages of this method
there is no cold start .
There is no sparsity.
Able to recomend to users with unique tests.
As a customer, you may experience AI which collaborates with recommender systems ,for example:
- AI use content based filtering by using Personalized product recommendations based on previous shopping activity or customer profile.
- Pricing optimization based on supply, demand, or previous shopping activity.
Recommender systems are a type of AI that provide
personalized product or content suggestions based on user behavior and preferences.
Examples of recommender systems include Netflix's movie recommendations and
Amazon's product suggestions. AI is the backbone of these systems, allowing
them to learn and adapt to user behavior over time.
References
Comments
Post a Comment