# US 2009/0276457 A1 · Method and System for Maintaining Profiles of Information Channels

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2009/0276457 A1
- Document ID: US20090276457A1
- Status: Published application
- Kind: A1
- Application number: 12/111,972
- Filed: 2008-04-30
- Published: 2009-11-05
- Earliest claimed: 2008-04-30
- Assignees: International Business Machines Corporation
- Inventors: David Carmel; Haggai Roitman; Elad Yom-Tov
- Web page: https://patentgrep.com/patent/US20090276457A1

## CPC

- G06F 16/9535: Search customisation based on user profiles and personalisation
- H04L 69/14: Multichannel or multilink protocols
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

A method and system are provided for maintaining profiles of information channels available on the Web, wherein the information channels are accessed via pull-only protocols. The method includes monitoring one or more channels by a channel pull action at a monitoring rate, wherein the monitoring rate is determined for the one or more channels based on the number of update events in a previous time period. The method may optimally include filtering the update events in the time period by a novelty measure, wherein the filtering disregards events that do not include significant novel information. The monitoring rate is adapted based on reinforcement learning applying iterative learning rules over time.

## Claims

### Claim 1

A method for maintaining profiles of information channels available on the Web, wherein the information channels are accessed via pull-only protocols, the method comprising:
monitoring one or more channels by a channel pull action at a monitoring rate, wherein the monitoring rate is determined for the one or more channels based on the number of update events in a previous time period.

### Claim 2 (depends on 1)

The method as claimed in claim 1, including filtering the update events in the time period by a novelty measure, wherein the filtering disregards events that do not include significant novel information.

### Claim 3 (depends on 1)

The method as claimed in claim 1, wherein the monitoring rate is adapted based on reinforcement learning applying iterative learning rules over time.

### Claim 4 (depends on 2)

The method as claimed in claim 2, wherein the filtering determines if update events include significant novel information by evaluating an update event's influence on channel cohesion, and disregarding update events with high similarity to an information channel's current profile.

### Claim 5 (depends on 4)

The method as claimed in claim 4, including setting a threshold parameter to control the amount of novelty filtering.

### Claim 6 (depends on 5)

The method as claimed in claim 5, wherein the threshold parameter is a similarity of an update event to the information channel's current profile, below which update events are disregarded.

### Claim 7 (depends on 1)

The method as claimed in claim 1, including monitoring multiple channels, wherein the monitoring rate is refined based on inter-channel profile similarities.

### Claim 8 (depends on 7)

The method as claimed in claim 7, including maintaining a similarity matrix with entries corresponding to the average profile similarity between two channels during a given time period.

### Claim 9 (depends on 1)

The method as claimed in claim 1, including monitoring multiple channels, wherein the M most important channels are monitored at any moment in time, wherein the importance of a channel is determined as a channel with a higher monitoring rate.

### Claim 10 (depends on 1)

The method as claimed in claim 1, including monitoring multiple channels, wherein M randomly selected channels are monitored at any moment in time.

### Claim 11 (depends on 1)

The method as claimed in claim 1, wherein a length of the previous time period in which the update events are monitored is varied to optimize the quality of results against resource use.

### Claim 12

A computer software product for maintaining profiles of information channels available on the Web, wherein the information channels are accessed via pull-only protocols, the product comprising a computer-readable storage medium, storing a computer in which program comprising computer-executable instructions are stored, which instructions, when read executed by a computer, perform the following steps:
monitoring one or more channels by a channel pull action at a monitoring rate, wherein the monitoring rate is determined for the one or more channels based on the number of update events in a previous time period.

### Claim 13

A method of providing a service to a client over a network for maintaining profiles of information channels available on the Web, wherein the information channels are accessed via pull-only protocols, the service comprising:
monitoring one or more channels by a channel pull action at a monitoring rate, wherein the monitoring rate is determined for the one or more channels based on the number of update events in a previous time period.

### Claim 14

A system for maintaining profiles of information channels available on the Web, wherein the information channels are accessed via pull-only protocols, the system comprising:
means for determining a monitoring rate for one or more channels based on a number of update events in a previous time period; and
means for monitoring one or more channels by a channel pull action at a monitoring rate.

### Claim 15 (depends on 14)

The system as claimed in claim 14, including a novelty filter to filter the update events in the time period, wherein the novelty filter disregards events that do not include significant novel information.

### Claim 16 (depends on 14)

The system as claimed in claim 14, wherein the means for determining a monitoring rate includes a reinforcement learning mechanism to apply iterative learning rules over time.

### Claim 17 (depends on 15)

The system as claimed in claim 15, including means for setting a threshold parameter to control the amount of novelty filtering.

### Claim 18 (depends on 14)

The system as claimed in claim 14, wherein the means for monitoring includes monitoring multiple channels, wherein the means for determining the monitoring rate includes means for refinement based on inter-channel profile similarities.

### Claim 19 (depends on 14)

The system as claimed in claim 14, wherein the means for monitoring includes monitoring multiple channels, and wherein the M most important channels are monitored at any moment in time, wherein the importance of a channel is determined as a channel with a higher monitoring rate.

### Claim 20 (depends on 14)

The system as claimed in claim 14, wherein the means for monitoring includes monitoring multiple channels, wherein M randomly selected channels are monitored at any moment in time.

### Claim 21 (depends on 14)

The system as claimed in claim 14, including means for varying the length of the previous time period in which the update events are monitored to optimize the quality of results against resource use.
