A Two Day Workshop on Data Mining

I recently attended this, 2 day workshop, July 30 and 31, 2010

Organized by Dept. of Computer Science & Engineering,

Reva Institute of Technology & Management, Bangalore.

• Chief Guest: Dr. M K Banga, Senior Manager, Wipro.
• Guest of Honour: Basawraj Patil, IBM Project Manager.
• Prof  & Head – VijayKumar, Reva Institute.
• Principal – Reva : Rana Pratap Reddy.

All awesome sessions:

Day - 1

Few Highlights:

• Data + Mine = DataMining
• Film = “Beautiful Minds”

Current Problems

• Data mining on Images / Audio / Video.
• Patterns, Similarity and Regularity.
• Mathematical Models, Statistics, Probability, Fuzzy Logic, Neural Networks
• Developing Mathematical models for Leaf pattern / Sun Flower
• Collaborative Filtering
• Stanford University offers – Dataset + Infrastructure
• 90% data on Internet is junk

• So, first Clean the data, Pre-processing steps [Data Cleansing and Transportation]
• Data Types: Enterprise Data / RDBMS / Web Data
• Book Reference: - “Pattern Classification” by Duda & Hart. 

Prof & Head - ISE, K. Raghuveer

raghunie@yahoo.com

National Institute of Engg, Mysore

• Supervised Learning/ Unsupervised Learning
• Classification / Clustering / Association
• Data Mining Tools:
• GLUTO – Graphical CLUstering Toolkit
• WEKA – Waikato Environment Knowledge Analysis [written in Java]
• XLMiner – Addin for Excel sheet.

      3 types of Web Mining:

        1) Web Content Mining

        2) Web Structure Mining

        3) Web Usage Mining

   

• Web Crawlers / Robot:
• Semantic Web
• IBM’s Clever Search Engine
• Extraction - Prepare a report of 1 page from 100 pages of Report.
• Abstraction – Prepare s summary/abstract from 100 pages of Report.
• Generate a title – Given 100 page project Report.

Srinivas Gowda
Advissory S/W Engineer
India Software Labs

• Data Quality – Critical factor to create consistent trust worthy information
• Association Rule: Market Basket Analysis
• Time Series Data – Mining ATM Applications
• SPSS – Statistical Package for Social Sciences

Day - 2

Few Highlights:

Distributed Data Mining

Clustering Techniques

vijaykbp@yahoo.co.in

kgsrinivas@msrit.edu

Dr. Srinivas K G

MSRIT, Bangalore

Support Vector Machines & Its Applications

n-p hard problem

Soft Computing

Tools – SPSS,WEKA

Stipulus Correlation – Correlating anything to anything

Data is fuzzy , No data is measurable

Fuzzy Dataset – Ex: She is beautiful, He is tall.

Local Data to Mine - Breast Cancer

SVM - Support Vector Machines

Retrieving interesting content has become a very difficult task.

Web Operating System
Virtual Society

Online Plagiarism Detector:

EDIA, EDAS and EasyChair

[% of text being copied more than 70%]

Challenges

• DMQL – Not a standard like SQL
• Scaling up for high dimensional data
• Sequential/ Time series Data
• Mining Complex Knowledge from Complex Data
• Impact Factor – Who all referred your paper
• Google Scholar
• Data Mining in Social Networking site
• Distributed Data Mining and Multi agent data
• Data Mining for Biological & environmental Problems
• Dealing with non-static, unbalanced and sensitive data
• UCI – Datasets – small & highly unbalanced

Prof Jay Bharatheesh Simha Ph.D
Abiba Systems

• Mining Massive Data Sets
• Data Mining for Knowledge Discovery
• BioInformitcis