Recommendation engine for upselling in restaurant orders
A computer-implemented method is provided to support a food ordering system for food items from a menu of a restaurant using natural language. Expressions made for ordering are used to recommend a food item that a user has a high probability of wanting to include in an order. The recommendation engine is trained using machine learning. Expressions are collected and parsed to identify words that might indicate food items offered by the restaurant. The words are provided to a restaurant owner to identify food items on a menu, with which the words are associated.
1 . A computer-implemented method comprising:
collecting spoken natural language expressions of a multiplicity of users of a food ordering system for a restaurant, the restaurant offering a menu of food items;
parsing the spoken natural language expressions to identify words being menu items to be included in the order and on the menu;
computing numbers of expressions using the identified words;
informing a system designer or restaurant owner of at least one identified word, which is unknown, from the identified words, wherein the at least one identified word appears with a predetermined ratio relative to identified words;
receiving, from the system designer or restaurant owner, an identification of an item on the menu that is to be associated with the at least one identified word; and
configuring the food ordering system to recognize the at least one identified word as a reference to the item so that the at least one identified word that is unknown is associated with the item and can be parsed and as a known identified word that is part of the items on the menu.
2 . The method of claim 1 , wherein identifying words being menu items includes recognizing at least one of acoustic and phonetic attributes of the natural language expressions to detect information about the natural language expressions in order to determine recommendations.
3 . The method of claim 1 , wherein identifying words being menu items includes searching for the words in a synonym database and searching for synonyms in descriptions of the food items.
4 . A food ordering system for a restaurant, the system comprising a processor and non-transitory memory, which stored code, wherein the processor executes the stored code and causes the food ordering system to:
parse, using a parser, spoken natural language expressions, which are collected from a multiplicity of users of the system to determine if the natural language expressions are items to be included in an order, to identify unknown words relative to menu items to produce identified words;
compute numbers of expressions of the identified words;
store in memory at least one of the identified words, wherein the at least one of the identified words appears with a predetermined ratio relative to other identified words;
identify a menu item on the menu associated with the at least one identified word; and
modify the system to recognize the at least one identified word as a reference to the menu item so that future unknown words are associated with menu items on the menu.
5 . A computer-implemented method comprising:
collecting spoken natural language expressions of a multiplicity of users of a food ordering system for a restaurant, the restaurant offering a menu of food items;
parsing the spoken natural language expressions using a language model to identify words being menu items to be included in the order and on the menu;
computing numbers of expressions using the identified words;
informing a system designer or restaurant owner of at least one identified word, which is unknown, from the identified words, wherein the at least one identified word appears with a predetermined ratio relative to identified words;
receiving, from the system designer or restaurant owner, an identification of an item on the menu that is to be associated with the at least one identified word; and
updating the language model to recognize the at least one identified word as a reference to the item so that the at least one identified word that is unknown is associated with the item and can be parsed by the language model as a known identified word that is part of the items on the menu.
6 . The method of claim 5 , wherein identifying words being menu items includes recognizing at least one of acoustic and phonetic attributes of the natural language expressions to detect information about the natural language expressions in order to determine recommendations.
7 . The method of claim 5 , wherein identifying words being menu items includes searching for the words in a synonym database and searching for synonyms in descriptions of the food items.