Subjective Search Intent Predictions using Customer Reviews
| Autor: | Adam Kiezun, Shay Artzi, Emily Dutile, Adrian Boteanu |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
050101 languages & linguistics
Information retrieval Computer science InformationSystems_INFORMATIONSTORAGEANDRETRIEVAL 05 social sciences Customer reviews Target audience 02 engineering and technology Search intent Product reviews Component (UML) 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing 0501 psychology and cognitive sciences Relevance (information retrieval) Product (category theory) |
| Zdroj: | CHIIR |
| Popis: | Query intent prediction is a component of information retrieval which improves result relevance through an understanding of latent user intents in addition to explicit query keywords. We target context-of-use intents, such as the activity for which a product is used and the target audience for a product, which are subjective and not usually indexed as product attributes in the catalog. We describe a method to predict latent query intents: we extract intents from product reviews on amazon.com and, using behavioral purchase signals that associate queries with the reviewed products, train query classifiers that label queries with the intents extracted from reviews. For example, we predict the activity "running" for the query "adidas mens pants." We show that our method can predict latent intents not indexed directly in the product catalog. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|