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 Detecting Trending Stocks from User Browsing Data

IP.com Disclosure Number: IPCOM000246133D
Publication Date: 2016-May-11
Document File: 2 page(s) / 18K

Publishing Venue

The IP.com Prior Art Database

Related People

Alina Beygelzimer: INVENTOR [+5]

Abstract

A method and system is disclosed for detecting trending stocks from user browsing data. The method and system serves as a stock discovery tool which discovers and ranks currently trending stocks based on user browsing activities that allow one or more users to learn about emerging stocks.

This text was extracted from a Microsoft Word document.
This is the abbreviated version, containing approximately 52% of the total text.

Method and System for Detecting Trending Stocks from User Browsing Data

Abstract

A method and system is disclosed for detecting trending stocks from user browsing data.  The method and system serves as a stock discovery tool which discovers and ranks currently trending stocks based on user browsing activities that allow one or more users to learn about emerging stocks.

Description

Disclosed is a method and system for detecting trending stocks from user browsing data.  The method and system serves as a stock discovery tool which discovers and ranks currently trending stocks based on user browsing activities that allow one or more users to learn about emerging stocks.  Further, the method and system increases user engagement by providing statistically meaningful and recent trends in stock browsing activities, ranking stocks by the significance of surge in the browsing activity.

In one implementation, the method and system detects trending stocks based on statistical testing.  The statistical testing is easy to implement and produces statistically valid results, however relies on being able to robustly estimate baseline frequencies.  For each stock s, the underlying data-generating process follows the binomial distribution – each click is either on s or not.  Let p be the baseline frequency with which clicks on s typically occurs.  The baseline frequency p can be estimated from past data.

Accordingly, for every sliding window, for every stock, the method and system computes the empirical frequency p using past window of data and computes a binomial upper confidence bound u on the baseline frequency.  Let N be the total number of clicks in the current window, out of which K clicks are on s. The expected number of clicks on stock s is pN < uN.  For K> uN, the right-tail p-value or the probability that X>=K, where X is a random variable that follows a binomial distribution with parameters N and u.  The stocks can be ranked by the...