• Ingredient product recommendations

Ingredient product recommendations

For a retailer we developed a crowdsourcing application that helps to assign appropriate product recommendations to recipe ingredients.

Human shoppers readily identify the appropriate retail products that are needed for a given recipe, but for computers this is a challenging task as ingredients may correspond to any of over a hundred of available products in the retailer’s assortment.

Our application uses natural language processing techniques to identify products in the product ontology that match novel ingredient descriptions. The application presents pictures of these products for each ingredient and solicits for product suggestions from the user. This information can later be used to automatically select or suggest products to online shoppers that wish to order the ingredients required for a selected recipe.

Similar crowdsourcing approaches can be used to address the cold start problem for recommendation systems in other domains.

More information?

Do you want more information about our crowdsourcing application or other Data Science solutions? Fill in the form below and we'll contact you. You can also contact us via 070-3195000