Personalized News Feed

To ensure the articles in the feed are fresh, you can restrict the recommendations to articles published within a certain timeframe (e.g., the last 3 days):

1. Navigate to the Filters tab and Add first filter.
2. Find Recent items, and click +.
3. Specify the property in your item catalog that represents the publish date.
4. Define the maximum age of articles (e.g., 3 days).

If the homepage includes other content sections that are not powered by Recombee, such as editor-picked articles, and you want to exclude these from the recommended feed, you can do so dynamically using the filter parameter in the API request.

For example, if you want to exclude articles with the IDs article-42, article-77, and article-92, you can use the following filter:

'itemId' not in {"article-42", "article-77", "article-92"}

Personalized feeds can also be applied to category pages. For example, if the user navigates to the "Sports" category, only articles related to sports should be displayed in the feed.

Since the selected category is dynamic, you should send the category filter as part of the filter parameter in the API request.

Depending on how you model the category/categories in your catalog, the filter (ReQL expression) may look like this:

String Property

'category' == "<the particular category>"

'category' == "sport"

Set of Categories

"<the particular category>" in 'categories'

"sport" in 'categories'

Alternative Implementation:

As an alternative to using a filter, you can utilize the Recommend Items to Item Segment endpoint for a specific category (= Item Segment). In this case, you don't need to provide a dynamic filter, as it is inferred from the given Item Segment.

To load subsequent articles when the user scrolls, use the Recommend Next Items endpoint in the same way as with Recommend Items to User.

If editors mark certain articles as preferred, and this information is exported to the Items catalog as a property, you can set up a Booster to increase the visibility of these articles.

1. Navigate to the Boosters tab and Add first booster.
2. Find Boost editors' picks and click +.
3. Specify the property in your item catalog that identifies the picked articles (the property should have a value of true for these articles).
4. Use the Preview tool to adjust the Boosting coefficient slider and find the optimal level of boosting for these articles.

If you want the personalized feed to always display articles from multiple topics or categories, follow these steps:

1. Navigate to the Constraints tab and Create Constraint.
2. Select the Segmentation that represents the topics or categories. If you don't have such Item Segmentation created yet, click the Create button and define the Segmentation by selecting the item property that contains the topic/category information.
3. Choose the Limit Type:
   • Absolute: Specify a fixed limit (e.g., allow at most 3 articles from a single topic per request).
   • Relative: Specify a percentage limit (e.g., ensure no more than 50% of the recommendations are from one topic).