Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Method and System for Context-based Mobile App Vectorization

IP.com Disclosure Number: IPCOM000246054D
Publication Date: 2016-Apr-29
Document File: 2 page(s) / 18K

Publishing Venue

The IP.com Prior Art Database

Related People

Qiang Ma: INVENTOR [+4]

Abstract

Disclosed is a method and system for modelling apps contextual similarity based on users' app usage activities. Based on usage context, a latent vector representation of each vector is derived. Thereafter, distance between vectorizations of the apps is used to measure similarity.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Method and System for Context-based Mobile App Vectorization

Abstract

Disclosed is a method and system for modelling apps contextual similarity based on users’ app usage activities.  Based on usage context, a latent vector representation of each vector is derived.  Thereafter, distance between vectorizations of the apps is used to measure similarity.

Description

There is a need to measure for app diversity and app importance with respect to a given user.  An importance of an app can be derived by measuring how similar the app is to one or more apps already installed by the user.  Accordingly, for measuring importance as well diversity, it becomes important to measure how similar two apps are. 

Therefore, given a set of users, and historic app usage sessions  of a user , where each app session  is represented by  where user  and app starting at time  and ending time , it is required to find similarity function  for two apps  and .

Existing solutions such as bag-of-words method and latent vectorization of apps for finding similarity do not take app usage context into account.  Similar apps that have different descriptions or complementary apps which are used in similar contexts but have different functionalities are not considered/learned as similar by the existing approaches.

Disclosed is a method and system for modelling apps contextual similarity based on users’ app usage activities.  Based on usage context, a latent vector representation of each vector is derived.  Thereafter, distance between vectorizations of the apps is used to measure similarity.

In accordance with the method disclosed herein, each app is treated as a word and each user is treated as a document consisting of app sessions that take place in order of time.

Consider the following app session sequence of a user:  (A, t1, A, t2, A, t3, B, t4, C, t5, A), where A, B, C are different apps, tis the elapsed time between two app sessions....