Friday, December 09, 2005

Machine Learning Topics

Particle Swarm Optimization
Swarm Intelligence

Ant Algorithms
ant colony optimization

Reinforcement Learning
Q-learning definition
Markov decision process

Computational Learning Theory
VC dimension

Principle of maximum entropy

Ensembles, Bagging and Boosting

Christophe Giraud-Carrier

Hidden Markov model

Saturday, October 29, 2005

Viral Marketing

Dr. Ralph F. Wilson suggests that Viral Marketing is comprised of the following components:
1. Gives away products or services
2. Provides for effortless transfer to others
3. Scales easily from small to very large
4. Exploits common motivations and behaviors
5. Utilizes existing communication networks
6. Takes advantage of others' resources
The effects of word-of-mouth, or viral marketing are motivations for utilizing the social network that customers belong in.

Friday, September 16, 2005

Customer Segmentation

Customer analysis helps a business better meet customer needs. Learning more about your customers is often benefited by intelligent segmentation. Customers can be segmented into a variety of groups. These segments can be based on behavioural, demographic, geographic, and psychographic variables. In fact customers can be segmented by any combination of these variables. Through viewing customers within such segments the problem of identifying and serving customers is simplifed. The knowledge provided by these segments is usually useful for determining actionable marketing tactics.

Tuesday, September 13, 2005

Stanford Data Mining Course

Stanford offers a nice Data Mining and Electronic Business course within the Statistics department. It looks like it covers many exciting aspects of the field.

Thursday, July 28, 2005

What is Lift?

In data mining, "lift" is often used to measure model performance. Here is a link to an article that explains how it is used: DMReview article

Wednesday, July 20, 2005

IP Country Lookup Tool

Here is a link to a tool that I created to lookup the countries for all of the IP addresses in a mess of text.

Thursday, June 30, 2005

Idea: Transaction Logger

It would be great to create a device that logs all of my transactions (whatever the method used to purchase) that could be carried around while shopping. This would then enable consumers to do personal data-mining on all of their own transactions. This would be a unique tool to help consumers improve purchasing habits and make smarter decisions. Additional interesting product associations could also be calculated and analyzed.

Friday, June 17, 2005

Web Data Mining (for Business Intelligence)

Bamshad Mobasher teaches a nice Web Mining course entitled "Web Data Mining (for Business Intelligence)" at DePaul University in Illinois. Currently, it is one of the few courses dedicated solely to this topic. I expect, as time goes, the number of courses on this topic will grow dramatically.

Thursday, June 02, 2005

Web Content Mining: Bing Liu

Bing Liu from the University of Chicago is very interested in Web Content Mining. He compiled of list of references regarding the topic. In addition, he gave a tutorial on Web content mining in Chiba, Japan in May, 2005.

Thursday, April 21, 2005

Natural Language Processing

Here is a nice introduction and dictionary for Natural Langauge Processing (NLP). This reference might come in handy when mining text documents.

Tuesday, March 08, 2005

Personalized Assistance System

A researcher, from Penn State, has been working on a personalized assistance system that automatically helps users find more relevant search results (see the article).

Wednesday, March 02, 2005

Stages of Knowledge Discovery in Websites

|--------------------| 3. PERSONALIZATION
|---------------------| 2. Advanced Web Mining
| 1. Clickstream Analysis

Personalization Companies

Rule-based Personalization:
ATG, BroadVision, Epiphany, Blue Martini

Collaborative Filtering Personalization:

Statistical Modeling:
Touch Clarity

State-based Personalization:

Monday, February 07, 2005

Social Network Analysis

  • The Oracle of Bacon
  • Erdös Number Project
  • Six Degrees of Separation Experiment
  • ReferralWeb

  • Scale-Free Networks

    Most social networks including the Web seem to be scale-free networks. A Scale-free network is unique because a small number of nodes are highly connected (e.g., the majority of nodes are sparsely connected).
  • Wikipedia Encyclopedia
  • Tuesday, January 25, 2005

    24 Key Database Marketing Techniques

    This article is a good overview of 24 marketing methods which are useful in determining what data should be collected in order to personalize a website.

    Wednesday, January 19, 2005

    Thursday, January 06, 2005

    Knowledge Discovery Approaches

    Data Mining enables us to automatically sift through mass amounts of data to discover KNOWLEDGE. A couple popular approaches are summarized below:

  • Identify customer groups and forecast their behaviour. This is commonly used in Marketing, fraud detection, and more and more frequently on the web for various purposes.

  • Market basket analysis: "If customer bought product P, he or she
    is likely to buy product Q and R" ( uses this approach)