Back to results
US 9,805,022 B2Granted
Generation of topic-based language models for an app search engine
AssigneeApple Inc.
InventorsNatalia Hernandez Gardiol, Catherine Anne Edwards
Life of this patent
Estimating the term…
Abstract
Topic-based language models for an application search engine enable a user to search for an application based on the application's function rather than title. To enable a search based on function, information is gathered and processed, including application names, descriptions and external information. Processing the information includes filtering the information, generating a topic model and supplementing the topic model with additional information.
Claims37
A method programmed in a memory of one or more devices comprising:
- a. analyzing a description of a mobile device application in a set of mobile device applications on one or more application stores;
- b. processing the description of the mobile device application into processed information by:
- grouping words frequently found together into topics and, for each topic, providing a topic probability for the topic that corresponds to a frequency with which the topic appears in the description of the mobile device application;
- and
- for each word in each topic, calculating a value for the word based on a probability that the topic contains the word;
- and
- c. generating a topic-based language model for a mobile device application search from the processed information;
- d. post-processing the processed information, the post-processing comprising comparing words in the topic-based language model with an application store application description and, when one or more words in the topic-based language model are not included in the app store application description, removing the one or more words from the topic-based language model;
- identifying query terms relevant to the mobile device application from log data that shows what users select after a search and adding one or more identified query terms to the topic-based language model based on the log data.
6 more elements
A non transitory computer readable medium comprising instructions encoded thereon for performing the method:
- a. extracting a first description of a mobile device application from an app store;
- b. crawling for a second description of the mobile device application on non-app store sources;
- c. concatenating the first description and the second description to form a combined description;
- d. filtering the combined description;
- and
- e. generating a topic model by:
- grouping words frequently found together into topics and, for each topic, providing a topic probability for the topic that corresponds to a frequency with which the topic appears in the description of the mobile device application;
- and
- for each word in each topic, calculating a value for the word based on a probability that the topic contains the word and a probability that a document in the description contains the topic
- f. post-processing the topic model, wherein post-processing the topic model comprises comparing words in the topic model with an app store application description and, when one or more words in the topic model are not included in the app store application description, removing the one or more words from the topic model;
- identifying query terms relevant to the mobile device application from log data that shows what users select after a search and adding one or more identified query terms to the topic-based language model based on the log data.
8 more elements
A device comprising:
- a. a memory for storing a program, the program:
- i. analyzing a first description and a second description of a mobile device application, the first description including an application name and an app store application description, the second description including a non-app store web source data;
- ii. combining the first description and the second description into a combined description;
- iii. filtering the combined description by removing non-descriptive language;
- iv. generating a topic model by:
- grouping words frequently found together into topics and, for each topic, providing a topic probability for the topic that corresponds to a frequency with which the topic appears in the description of the mobile device application;
- and
- for each word in each topic, calculating a value for the word based on a probability that the topic contains the word and a probability that a document in the description contains the topic;
- and
- v. supplementing the topic model by performing post-processing, the post-processing comprising comparing words in the topic model with an app store application description and, when one or more words in the topic model are not included in the app store application description, removing the one or more words from the topic model;
- and
- b. a processing component coupled to the memory, the processing component configured for processing the program;
- identifying query terms relevant to the mobile device application from log data that shows what users select after a search and adding one or more identified query terms to the topic-based language model based on the log data.
10 more elements
Dependencies are inferred from the claim text.