Web

April 2017
Business Economics & Information Technology

OUTLINE
◼Web mining
◼Data mining/Data mining techniques/ Data mining Algorithms
◼Social media mining
◼Text mining
◼Categories of web mining
Web content mining
Web Usage Mining
Web Structure Mining
https://orange.biolab.si/

WHAT IS WEB MINING?
Web Mining is the use of the data mining techniques to automatically discover and
extract information from web.
Web Mining can find interesting and potentially useful knowledge from web data

WHAT IS DATA MINING?
Data mining or knowledge discovery from data is the process of analyzing data from
different perspectives and summarizing it into useful information
Knowledge Discovery in Databases
Raw data knowledge

DATA MINING TECHNIQUES
 Clustering
 Classification
 Association Rules
 Correlation
 Naive Bayesian
 Neural Networks
 Outlier detection/ Anomaly detection
 Regression
 Logistic Regression
The most popular data mining techniques are:

WHAT IS WEB DATA?
Web content –text , image, records, etc.
 Web structure – hyperlinks, tags, etc.
 Web usage –http log , app server logs ,etc
Intra-page structures- document level
 Inter-page structures- hyperlink level
 Supplemental data
 Profiles
 Registration information
Cookies

DATA MINING VS. WEB MINING
Data Mining
Data is structured and relational
Well-defined tables, columns, rows, keys, and constraints.
Web Mining
Semi-structured(HTML) and unstructured

EXAMPLE: ASSESSING CREDIT RISK
Situation: Person applies for a loan
Task: Should a bank approve the loan?
Note: People who have the best credit don’t need
the loans, and people with worst credit are not likely to repay.
Bank’s best customers are in the middle.

EXAMPLE: INSURANCE FRAUD
Insurance Fraud is the filing of a false claim to life, health, automobile, property or
other types of insurance benefits.
Insurance companies lose millions of dollars each year through fraudulent claims,
largely because they do not have a way to easily determine which claims are legitimate
and which may be fraudulent.

EXAMPLE: INSURANCE FRAUD
Data mining enables insurance companies to predict which insurance claims are likely
to be fraudulent.
http://www.hugin.com/solutions/fraud-detection-management/online-demonstration

OPPORTUNITIES & CHALLENGES
 The amount of information on the Web is huge
 The coverage of Web information is very wide and diverse.One can Find information
about almost anything. Information/data of almost all types exist on the Web. For
example, structured tables, texts, stream data, etc.
Much of the Web information is semi-structured due to the nested structure of HTML
code.
 Much of the Web information is linked. There are hyperlinks among pages within a
site, and across different sites.
 Much of the Web information is redundant. The same piece of information or its
variants may appear in many pages.

OPPORTUNITIES & CHALLENGES
The Web is noisy.A Webpage generally contains a mixture of many kinds of
information. For example: main contents, advertisements, navigation panels, copyright
notices, etc.
The Web is dynamic. New pages are constantly being generated. Keeping up with the
changes and monitoring the changes are important issues.
Above all, the Web is a virtual society. It is not only about data, information and
services, but also about interactions among people, organizations and automatic
systems,and communities.

APPLICATION OF WEB MINING IN E-COMMERECE
Customer Analyzing
Mined data help acquire new, retain existing customers, Improvement of merchant services and
profit by predicting customer online purchase behavior
◼What do the customers do?
◼What do the customers want?
◼How effectively use the web data to market products and to service the customer?
◼Whether customers are purposefully or just browsing?
◼Buying something they are familer with or something they know little about?
◼Are they shopping from home, from work or from a hotel?

Web personalization
According to the information from user behavior, a website can be designed and re-structured to
make it more advance and user-friendly. In addition, the image and product value of the
company is very important in satisfying customer need based on website quality.
Personalizing a website involves tailoring content based on the characteristics of each
individual user’s online behaviors.
Personalized content is often determined by user behaviors such as pages viewed, buttons
clicked and forms submitted.

Product search & Recommendation
When the user searches for a product how we find the best results for the users?
Typically, a user query of a few keywords can match many products.
 Through large-scale data analysis of query logs, we can create graphs between queries and products, and
between different products.
 For example, the user who searches for “Verizon cell phones” might click on the Samsung SCH U940 Glyde
product, and the LG VX10000 Voyager. We now know the query is related to those two products, and the two
products have a relationship to each other since a user viewed (and perhaps considered buying) both.

CATEGORIES OF WEB MINING
Web mining is divided into three categories:
1.Web Content Mining
2. Web Usage Mining
3. Web Structure Mining

WEB CONTENT MINING
To gather, categorize, organize and provide the best possible information available on the web to the user
requesting the information
The data may be unstructured or structured (data from a database) or semi-structured (html)
Content mining is the scanning and mining of text, pictures, video, audio and graphs of a Web page to
determine the relevance of the content to the search query
Content mining provides the results lists to search engines in order of highest relevance to the keywords in
the query
Web content mining is related to data mining and text mining Discovering useful information
from contents of Webpages

TEXT MINING
Text mining is the analysis of data contained in natural language text
Text mining attempts to derive meaning from the words and sentences in order to
classify documents, route messages appropriately, as well as create summaries of
content
Unstructured Data Examples: Email, Insurance Claim,
Web Pages, Technical Documents, Contracts
 https://www.nytimes.com/2016/09/24/us/politics/presidential-debate-hillary-clinton-donald-trump.html?_r=0
 https://www.youtube.com/watch?v=Ozo2QuCKml0
https://voyant-tools.org/

DATA MINING TECHNIQUES USING IN WEB CONTENT MINING
The more basic and popular data mining techniques in web content mining are:
Classification : Placing the documents into a predefined set of groups such as science articles, Political
articles, etc.
Clustering : Clustering is a technique used to group similar documents (is not done based on
predefined). As a result useful documents will not be omitted from the search results. Clustering helps the
user to easily select the topic of interest.
Summarization is used to reduce the length of the document by maintaining the main points. An
example for text Summarization is Microsoft word’s AutoSummarize
Visualization utilizes feature extraction and key term indexing to build a graphical representation.
Through visualization, documents having similarity are found out is useful to find out related topic from a
very large amount of documents. Examples: Word Cloud, Scatter Plot, Streamgraph, Tree map, Heat map,
Gantt Chart, etc.

WEB USAGE MINING
Web usage mining
 Is used to understand the customer behavior
Focuses on the discovering of potential knowledge from browsing patterns of the users.
Can discover the knowledge in the hidden browsing patterns and analyses the visiting characteristics of the
users.
The primary data source used in web usage mining is the server log-files (web-logs).
Browsing web pages by the user leaves a lot of information in the log-file.
Analyzing log-files information drives us to understand the behavior of the user
Techniques use for discovering the potential knowledge from the browsing patterns are:
Clustering
Classification
Association rule
40% of Online Shopper don't complete
their purchases

THE PHASES OF WEB USAGE MINING

CLASSIFICATION
Classification is the most familiar and most popular data mining technique for web usage
mining.
Data classification is the process of organizing data into categories for its most effective and
efficient use.
Classification technique uses to segment and classify observations
Example :
People with age less than 40 and salary more than 40000, trade
on line(Demographic segmentation ) .
 Blackberry was launched for users who were business people, Samsung was launched for
users who like android and like various applications for a free price, and Apple was launched for
the premium customers who want to be a part of a unique and popular niche(Behavioral
segmentation)

CLASSIFICATION
Classification consist of assigning a class label to a set of unclassified cases.
The goal of classification is to build a model that can be used to predict the class of records whose class
label is not Know.

CLASSIFICATION ALGORITHMS
The most popular classification algorithms are:
Decision trees
Logistic regression
Neural networks
k-nearest neighbors

DECISION TREES
◼A decision tree is a graph that uses a branching method to illustrate every possible outcome of a decision.
EXAMPLE

DECISION TREES
Customer online-
shopping
behavior

Decision Tree using Orange Data Mining
Analysing data in Orange using Decision tree.
Select file: Decision tree from Dataset Folder(On Fronter)
Exercise:
Explain the output of the Decision tree

CLUSTERING
◼Clustering is the process of dividing a dataset into groups such that the members of
each group are as similar as possible to one another and different groups are as
dissimilar as possible from one another
◼The most popular distance-based clustering algorithms is ‘k-means’.

K MEANS FOR CLUSTERING
K-Means Algorithm for Clustering
The number of car accident is
classified by population

CLUSTERING USING ORANGE
Select file: Clustering from Dataset Folder(On Fronter)
Select K-Means from Unsupervised Widget set.
Select MDC(Multidimensional scaling )
Unsupervised Widget set
Exercise:
Explain the output of the Clustering
to create a segmentation based only on buying behavior
https://archive.ics.uci.edu/ml/datasets/Wholesale+customers

ASSOCIATION RULE
Association rule finds interesting associations and correlation
relationships among large sets of data items.
Association rules show attribute value conditions that occur frequently
together in a given data set.
A typical example of association rule mining is Market Basket Analysis.
What items are frequently
bought together by customers?

EXAMPLE OF MARKET BASKET
Items are frequently
bought together by customers, should be
placed together in the store to maximize
sales.

PRODUCT OFFER & RECOMMENDATIONS
IF {milk, flour, sugar, eggs, candles} THEN {party hats, paper plates, magician}

Association analysis in Orange
Select file: Association Rulefrom Dataset Folder(On Fronter)
Select Data Table from Data at the Widget set.
Select Frequent Itemset from Associate
Select Association Rules from Associ
Exercise:
Explain the output of the Association
https://www.lynda.com/Business-Intelligence-tutorials/Association-analysis-
Orange/475936/529739-4.html

WEB STRUCTURE MINING
 The structure of a Web consists of Web pages as nodes, and hyperlinks as edges
connecting between two related pages
 The research at the hyperlink level is also called HYPERLINK
ANALYSIS
 Web structure mining is to study the relationship between the reference pages to find useful
patterns, and improve search quality by analyzing the links between pages
 Web structure Mining focuses on
Reducing irrelevant search results
Help indexing information on the web

Web Structure Terminology
Web-Graph: A directed graph that represent the web.
Node: Each Web page is a node of the Web-graph.
Link: Each hyperlink on the Web is a directed edge of the Web-graph.
In-degree: The in-degree of a node, p is the number of distinct links that
point to p.
Out-degree: The out-degree of a node, p is the number of distinct links
originating at p that point to other nodes.

Web Structure Terminology
Directed Path: A sequence of links, starting from p that can be followed to reach q.
Shortest Path: Of all the paths between nodes p and q, which has the shortest length, i.e.
number of links on it.
Diameter: The maximum of all the shortest paths between a pair of nodes p and q, for all pairs of
nodes p and q in the Web-graph (the length of the longest shortest path)

Hubs and authorities are ‘fans’ and ‘centers’ of a web graph
A good hub page is one that points to many good authority pages
A good authority page is one that is pointed to by many good hub pages
Hubs and Authorities

Google’s Page Rank
Rank of a web page depends on the rank of the web pages
pointing to it
Hyperlink analysis algorithm assigns numerical weight to a
webpage
Page Rank increases effectiveness of search engines
To Climb to The Top of Google Search

SOCIAL MEDIA MINING
Social media mining is the process of representing, analyzing, and extracting actionable patterns and trends
from raw social media data.
Social media mining uses a range of basic concepts from computer science, data mining, machine learning,
and statistics.
Social media mining is based on theory from social network analysis(SNA)
Data mining techniques in social media mining are:
Graph Mining
Text Mining

SOCIAL NETWORK ANALYSIS
Social network analysis [SNA] is the mapping and measuring of relationships and flows between
people, groups, organizations, computers, and other connected information/knowledge entities.
The nodes in the network are the people and groups while the links show relationships or flows
between the nodes.
 SNA provides both a visual and a mathematical analysis of human relationships.
EXAMPLE:
Who knows whom and who shares what information
and knowledge with whom through what media.

GRAPH MINING
Extracting useful knowledge (patterns, outliers, etc.) from structured data that can be represented as a grap
https://neo4j.com/download/
A Graph is a set of nodes and the
relationships that connect those nodes
 Nodes and Relationships contain
properties to represent data.

TEXT MINING
◼A social network contains a lot of data in the nodes of various forms. For example, a
social network may contain blogs, articles, messages, and etc.
◼ Common application for text mining is to aid in the automatic classification of texts.
For example, it is possible to "filter" out automatically most undesirable "junk email"
based on certain terms or words that are not likely to appear in legitimate messages

EXERCISE
Compare two competitive products in Social Media base on the comments on the products (Use text mining tool-Voyant)

SUMMARY
◼ Web mining
◼ Data mining
◼ Data mining techniques
◼ Web Data
◼ Applications of web mining in E-commerce
◼ Categories of web mining
 Web content mining
 Text mining
 Data mining
o Classification
o Clustering
o Summarization
o Visualization
 Web Usage Mining
 Clustering –K means algorithms
 Classification – Decision Tree
 Association rule –Basket Analysis
 Web Structure Mining
◼ Social Media Mining
 Graph Mining
 Text Mining

Web

More Related Content

Similar to Web (20)

Recently uploaded (20)

Web