Make your own free website on Tripod.com

COMP 578

Data Mining

Assignment # 1

Question 1.7 # 2:

Suppose that you have employed as a data mining consultant for an Internet search engine company. Describe how data mining can help the company by giving specific example of how techniques, such as clustering, classification, association rule mining, and anomaly detection can be applied.

  1. Classification: This is a predictive model.Classificationís main goal is to assign previously unseen records a class as accurately as possible.Usually training sets are used to build the model.Test sets are used to determine the accuracy of the model.Each training sets contains records.Each record contains different attributes.Each attribute is known as a class.One of the ways that a company can use classifications is to assist with predicting the target consumers of new products based on similar past products.This would assist with cost savings with targeted and direct marketing to those particular consumers most likely to purchase the product based on past historical records of consumerís spend patterns and other different attributes such as income, demographics, age, lifestyle etc.
  2. Clustering: This is a descriptive model.Data is clustered by similarities.Data points that are more similar to each other are clustered together.Each cluster differs from one another due to the less similar data points.One way that companies can use clustering is to target their support to certain segments of their customer base.By subdividing customers according to attributes such as buying patterns, spending amounts etc, different service levels can be offered by the company.This can assist companies to ensure that they provide the best support to their customers that spend the most with them.
  3. Association Rule Discovery: This is a descriptive model.This is used for pattern recognition.Records are associated with each other through dependency rules to predict an occurrence of an item given the occurrence of other items.An example for use of this method is suggestions made by online book-sellers such as Amazon.com.When one buys a book, association rule discovery is used to prompt users into buying other books that previous purchasers bought in addition to the original book.
  4. Anomaly Detection:This is a predictive model.This is where sets of values are analyzed for anomalies in the data sets.One of the most significant uses of this method is for Credit Card Fraud detection.Credit Card companies record the usage patterns of their customers.If there are anomalies in those usage patterns, this will be alerted and can be followed up by the company for further analysis.