Pin Concepts

Concept pinning is an interesting feature that allows you to classify your request tokens into a Boolean category. This may play an important role in your search engine query processing. A keyword based search engine performs a Boolean AND or OR operation on the query or they perform a Min Should Match (eDismax mm feature) query. The Pin Concepts feature can be used to tag the significant tokens in the query.The system ensures that the pinned tokens must be matched in the resultant documents.

The following are the steps to enable the Pinning in a pipeline:

  1. Go to the Advanced tab and select the Pin Concepts checkbox
  2. Select the categories whose tokens you want to mark as pinned.

In addition to Pinning concepts, you can also specify the concepts which should never be Pinned. These include those concepts or categories which should always be ‘optional’ in user query. This is done through ‘blacklisting’ some concepts from Pinning.

The following are the steps to blacklist concepts from Pinning:

  1. Go to the Advanced tab and select the Pin Concepts checkbox.
  2. Open the Blacklist configuration area by clicking on ‘+’ icon before ‘Blacklist’.
  3. Select all Vocabularies and/or Entities that you wish to Blacklist from pinning. You could also enter individual concepts in the ‘Terms’ area.
  4. Save the configurations.

Text Summarizer

This feature summarizes the provided paragraphs of text into a few sentences. It can be used for displaying a snippet of text in the search result page. It may not be very useful for information retrieval problem but is a crucial NLP feature.

The following are the steps to enable the text summarizer in a pipeline:

  1. Go to the Advanced tab and select the Text Summarizer checkbox.
  2. Specify the maximum number of sentences in which you want to summarize the text.

Sentiment Analysis

Sentiment Analysis detects the polarity of the specified text as negative, neutral or positive.

The following are the steps to enable the sentiment analysis in a pipeline:

  1. Go to the Advanced tab and select the Sentiment Analysis check box.