Browse Prior Art Database

Social Contact Ranking / Suggestion for Personal Messaging

IP.com Disclosure Number: IPCOM000191090D
Original Publication Date: 2009-Dec-15
Included in the Prior Art Database: 2009-Dec-15
Document File: 3 page(s) / 108K

Publishing Venue

IBM

Abstract

In an increasingly-connected world, it is easy to connect with a given party via IM, email, text, phone call, or other communication methods. Given a plethora of online directories and address books, however, it is also just as easy to inadvertently message or call the wrong party. Auto-complete solutions exist today that detect and suggest contact names based on the recency of contact (e.g. Sametime8 will suggest "John Smith / Austin" before "John Smith / San Jose" if I've recently IM'd with John Smith in Austin). However, in cases where the user has not communicated with the intended user, such schemes are unable to provide help. Given the large amounts of data available from online address books, social networks (e.g. Facebook), and IM contact lists, auto-complete software should be able to make intelligent suggestions based on the "closeness" of the possible target contacts, where "closeness" is defined by proximity in a social network or organizational chart.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 74% of the total text.

Page 1 of 3

Social Contact Ranking / Suggestion for Personal Messaging

A system may rank potential contacts by one of several possible methods:

Method 1.a. Detection of unusual messages or sorting of contacts (IMs, TXTs, emails, VOIP calls) based on a dynamically-built social network (built server-side).

    A collaboration system could assign a weighting value based on some closeness algorithm or strength of link between the sender, and the potential target user. If the connection is too tenuous, confidence that this is the intended recipient will be low. A further extension might rank users based on proximity, and popularity in the network (e.g. 4 possible paths between the sender and "recipient 1" would be ranked ahead of 1 possible path between the sender and "recipient 2."

Method 1.b. The same as the above, built on the client, after analyzing messages on the users' device, using that device's communications to build a real-time map of the sender's recent social contacts.

1

Page 2 of 3

Figure 1: Social Network Map, where a likely candidate "1" is selected ahead of candidates 2 and 3, based on proximity in the network.

Method 2. Detection of unusual messages, or sorting of based on organizational charts. Based on report-to chains, software may rank potential recipients based on the number of hops between the sender and the recipient in the organizational c...