Browse Prior Art Database

Method and System for Creating Mobile Based Lexicon based on Mobile Forecasting

IP.com Disclosure Number: IPCOM000247827D
Publication Date: 2016-Oct-06
Document File: 5 page(s) / 119K

Publishing Venue

The IP.com Prior Art Database

Abstract

A method and system is disclosed for enhancing accuracy of probabilistic systems by leveraging capabilities of a mobile device to capture the context of a user. A lexicon or a repository of words are constructed from a mobile device. The method and system derives and disambiguates semantic definitions and ranking for the words from mobile based forecasting.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 38% of the total text.

Page 01 of 5

Method and System for Creating Mobile Based Lexicon based on Mobile Forecasting

Mobile and untethered devices are increasing accessibility for probabilistic distributed natural language processing (NLP) systems. Contextual clues about integrated usage of mobile devices in a consumer's daily life provides semantic and lexical support for natural language systems. The accuracy of the probabilistic systems can increase while enabling better tradeoffs between precision, recall, selectivity, and sensitivity.

Disclosed is a method and system for enhancing accuracy of probabilistic systems by leveraging capabilities of a mobile device to capture the context of a user. A lexicon or a repository of words are constructed from a mobile device. The method and system derives and disambiguates semantic definitions and ranking for the words from mobile based forecasting.

In accordance with an embodiment, various components of the method and system are illustrated as follows.

The method and system includes a Mobile Forecaster that builds forecasting in order to answer a user's questions, a Mobile Lexicon Generator to produce and to generate lexical analyzers. A lexical analyzer is a program that transforms a stream of characters into a stream of 'atomic chunks of meaning' that are generated in conjunction with mobile forecasted data. Further the method and system include an NLP Mobile Context Dependent Analytical Pipeline to annotate the mobile forecasted data that creates machine learning features based on the question text, answer and the context in which answer appeared such as a passages or evidences. An NLP Mobile Context Dependent Forecaster Pipeline is also provided to annotate the mobile forecasted data that creates machine learning features based on the question text, answer without regard to the context of the answer. Further, the method and system includes a mobile machine learning that influence the performance achieved by existing learning systems and a Mobile Machine Learning Apply and Publishing to NLP System that applies machine learning into NLP system from mobile devices.

The method and system alters machine learning ensemble weights based on mobile forecasting based on, but not limited to, properties such as predictions based on user movement determined by mobile device sensors, features for forecasting based on recent multimedia capture from mobile devices, textual nugget capture for forecasting (e.g. summarization of recently read articles on a mobile device) and logging of most recently used applications.

The machine learning ensemble includes, but is not limited to, logistic regression phases and routes, dimensions of evidence refinement, probability density functions, classification techniques and regression techniques, NLP systems and any data retrieval systems. Further, the environment of the method and system can include, but need not be limited to, a plurality of mobile devices that work together, a single mobile devi...