Chapter I:Introduction MIS /2017 Fall

Chapter I:Introduction MIS 331 2016/2017 Fall

Chapter 1. Introduction Motivation: Why data mining?
Methodology of Knowledge Discovery in Databases Data mining functionalities Are all the patterns interesting? Business applications of data mining

Motivation: “Necessity is the Mother of Invention”
Data explosion problem Automated data collection tools and mature database technology lead to tremendous amounts of data stored in databases, data warehouses and other information repositories Need to convert such data into knowledge and information Applications Business management Production control Market analysis Engineering design Science exploration

Evolution of Database Technology (1)
Data collection, database creation Data management data storage and retrieval database transaction processing Data analysis and understanding Data mining and data warehousing

Evolution of Database Technology (2)
Data collection, database creation, IMS and network DBMS 1970s: Relational data model, relational DBMS implementation 1980s: RDBMS, advanced data models (extended-relational, OO, deductive, etc.) Application-oriented DBMS (spatial, scientific, engineering, etc.) 1990s: Data mining, data warehousing, multimedia databases, and Web databases 2000s Stream data management and mining Data mining and its applications Web technology (XML, data integration) and global information systems

The Explosive Growth of Data: from terabytes to petabytes
Data collection and data availability Automated data collection tools, database systems, Web, computerized society Major sources of abundant data Business: Web, e-commerce, transactions, stocks, … Science: Remote sensing, bioinformatics, scientific simulation, … Society and everyone: news, digital cameras, YouTube We are drowning in data, but starving for knowledge! “Necessity is the mother of invention”—Data mining—Automated analysis of massive data sets

Developments in computer hardware
Powerful and affordable computers Data collection equipment Storage media Communication and networking

Data Warehouse Data cleaning Data integration
OLAP: On-Line Analytical Processing summarization consolidation aggregation view information from different angles but additional data analysis tools are needed for classification clustering charecterization of data changing over time

Data rich information poor situation
Abundance of data need for powerful data analysis tools “data tombs” - data archives seldom visited Important decisions are made not on the information rich data stored in databases but on a decision maker’s intuition no tool to extract knowledge embedded in vast amounts of data Expert system technology domain experts to input knowledge time consuming and costly

What Is Data Mining? Data mining (knowledge discovery in databases):
Extraction of interesting (non-trivial, implicit, previously unknown and potentially useful) information or patterns from data in large databases Alternative names and their “inside stories”: Data mining: a misnomer? Knowledge discovery(mining) in databases (KDD), knowledge extraction, data/pattern analysis, data archeology, data dredging, information harvesting, business intelligence, etc. What is not data mining? query processing. Expert systems or small ML/statistical programs

Data Mining vs. Data Query
Data Query:e.g. A list of all customers who use a credit card to buy a PC A list of all MIS students having a GPA of 3.5 or higher and has studied 4 or less semesters Data Mining problems:e.g. What is the likelihood of a customer purchasing PC with credit card Given the characteristics of MIS students predict her SPA in the comming term What are the characteristics of MIS undergrad students

Why Data Mining? Four questions to be answered
Can the problem clearly be defined? Does potentially meaningful data exists? Does the data contain hidden knowledge or useful only for reporting purposes? Will the cost of processing the data will be less then the likely increase in profit from the knowledge gained from applying any data mining project

Steps of a KDD Process (1)
1. Goal identification: Define problem relevant prior knowledge and goals of application 2. Creating a target data set: data selection 3. Data preprocessing: (may take 60%-80% of effort!) removal of noise or outliers strategies for handling missing data fields accounting for time sequence information 4. Data reduction and transformation: Find useful features, dimensionality/variable reduction, invariant representation.

Steps of a KDD Process (2)
5. Data Mining: Choosing functions of data mining: summarization, classification, regression, association, clustering. Choosing the mining algorithm(s): which models or parameters Search for patterns of interest 6. Presentation and Evaluation: visualization, transformation, removing redundant patterns, etc. 7. Taking action: incorporating into the performance system documenting reporting to interested parties

An example: Customer Segmentation
1. Marketing department wants to perform a segmentation study on the customers of AE Company 2. Decide on revevant variables from a data warehouse on customers, sales, promotions Customers: name,ID,income,age,education,... Sales: hisory of sales Promotion: promotion types durations... 3. Hendle missing income, addresses.. determine outliers if any 4. Cenerate new index variables representing wealth of customers Wealth = a*income+b*#houses+c*#cars... Make neccesary transformations z scores so that some data mining algorithms work more efficiently

Example: Customer Segmentation cont.
5.a: Choose clustering as the data mining functionality as it is the natural one for a segmentation study so as to find group of customers with similar charecteristics 5.b: Choose a clustering algorithm K-means or k-medoids or any suitable one for that problem 5.c: Apply the algorithm Find clusters or segments 6. make reverse transformations, visualize the customer segments 7. present the results in the form of a report to the marketing deprtment İmplement the segmentation as part of a DSS so that it can be applied repeatedly at certain internvals as new customers arrive Develop marketing strategies for each segment

Data Mining: A KDD Process
Knowledge Pattern Evaluation Data mining: the core of knowledge discovery process. Data Mining Task-relevant Data Selection Data Warehouse Data Cleaning Data Integration Databases

Data Mining in Business Intelligence
Increasing potential to support business decisions End User Decision Making Data Presentation Business Analyst Visualization Techniques Data Mining Data Analyst Information Discovery Data Exploration Statistical Summary, Querying, and Reporting Data Preprocessing/Integration, Data Warehouses DBA Data Sources Paper, Files, Web documents, Scientific experiments, Database Systems October 23, 2017 19 Data Mining: Concepts and Techniques 19

Architecture of a Typical Data Mining System
Graphical user interface Pattern evaluation Data mining engine Knowledge-base Database or data warehouse server Filtering Data cleaning & data integration Data Warehouse Databases

Architecture of a Typical Data Mining System
Data base, data warehouse Data base or data warehouse server Knowledge base concept hierarchies user beliefs asses pattern’s interestingness other thresholds Data mining engine functional modules characterization, association, classification, cluster analysis, evolution and deviation analysis Pattern evaluation module Graphical user interface

Data Mining: Confluence of Multiple Disciplines
Database Technology Statistics Data Mining Machine Learning Visualization Information Science Other Disciplines

Why Confluence of Multiple Disciplines?
Tremendous amount of data Algorithms must be highly scalable to handle such as tera-bytes of data High-dimensionality of data Micro-array may have tens of thousands of dimensions High complexity of data Data streams and sensor data Time-series data, temporal data, sequence data Structure data, graphs, social networks and multi-linked data Heterogeneous databases and legacy databases Spatial, spatiotemporal, multimedia, text and Web data Software programs, scientific simulations New and sophisticated applications Add a definition/description of “traditional data analysis”. October 23, 2017 23 Data Mining: Concepts and Techniques 23

Efficient and Scalable Techniques
For an algorithm to be efficient and scalable its running time should be predictable and acceptable How Parallel and distributed algorithms Sampling from databases

Two Styles of Data Mining
Descriptive data mining characterize the general properties of the data in the database finds patterns in data and the user determines which ones are important Predictive data mining perform inference on the current data to make predictions we know what to predict Not mutually exclusive used together Descriptive  predictive Eg. Customer segmentation – descriptive by clustering Followed by a risk assignment model – predictive by ANN

Supervised vs. Unsupervised Learning
Supervised learning (classification, prediction) Supervision: The training data (observations, measurements, etc.) are accompanied by labels indicating the class of the observations New data is classified based on the training set Unsupervised learning (summarization. association, clustering) The class labels of training data is unknown Given a set of measurements, observations, etc. with the aim of establishing the existence of classes or clusters in the data

Descriptive Data Mining (1)
Discovering new patterns inside the data Used during the data exploration steps Typical questions answered by descriptive data mining what is in the data what does it look like are there any unusual patterns what dose the data suggest for customer segmentation users may have no idea which kind of patterns may be interesting

Descriptive Data Mining (2)
patterns at verious granularities geograph country - city - region - street student university - faculty - department - minor Fuctionalities of descriptive data mining Clustering Ex: customer segmentation summarization visualization Association Ex: market basket analysis

A model is a black box X: vector of independent variables or inputs
Y =f(X) : an unknown function Y: dependent variables or output a single variable or a vector Model Y output inputs X1,X2 The user does not care what the model is doing it is a black box interested in the accuracy of its predictions

Predictive Data Mining (1)
Using known examples the model is trained the unknown function is learned from data the more data with known outcomes is available the better the predictive power of the model Used to predict outcomes whose inputs are known but the output values are not realized yet Never %100 accurate

Predictive Data Mining (2)
The performance of a model on past data is not important to predict the known outcomes Its performance on unknown data is much more important

Typical questions answered by predictive models
Who is likely to respond to our next offer based on history of previous marketing campaigns Which customers are likely to leave in the next six months What transactions are likely to be fraudulent based on known examples of fraud What is the total amount spending of a customer in the next month

Data Mining Functionalities (1)
Concept description: Characterization and discrimination Generalize, summarize, and contrast data characteristics, e.g., big spenders vs. budget spenders Association (correlation and causality) Multi-dimensional vs. single-dimensional association age(X, “20..29”) ^ income(X, “20..29K”) à buys(X, “PC”) [support = 2%, confidence = 60%] contains(T, “computer”) à contains(x, “software”) [1%, 75%]

Classification and Prediction Finding models (functions) that describe and distinguish classes or concepts for future prediction E.g., classify people as healty or sick, or classify transactions as fraudulent or not Methods: decision-tree, classification rule, neural network Prediction: Predict some unknown or missing numerical values Cluster analysis Class label is unknown: Group data to form new classes, e.g., cluster customers of a retail company to learn about characteristics of different segments Clustering based on the principle: maximizing the intra-class similarity and minimizing the interclass similarity

Outlier analysis Outlier: a data object that does not comply with the general behavior of the data It can be considered as noise or exception but is quite useful in fraud detection, rare events analysis Trend and evolution analysis Trend and deviation: regression analysis Sequential pattern mining: click stream analysis Similarity-based analysis Other pattern-directed or statistical analyses

Concept Description Characterization Discerimination Data
classes or concpets classes of items for sale computers, printers concepts of customers: bigSpenders BudgetSpenders

Data Characterization
Summarization the data of the class under study (target class) Methods SQL queries OLAP roll up -operation user-controlled data summarization along a specified dimension attribute oriented induction without step by step user interraction the output of characterization pie charts, bar chars, curves, multidimensional data cube, or cross tabs in rule form as characteristic rules

Characterization example
Description summarizing the characteristics of customers who spend more than $1000 a year at AllElecronics age, employment, income drill down on any dimension on occupation view these according to their type of employment

Data Discrimination Comparing the target class with one or a set of comparative classes (contrasting classes) these classes can be specified by the use database queries methods and output similar to those used for characterization include comparative measures to distinguish between the target and contrasting classes

Discrimination examples
Example 1:Compare the general features of software products whose sales increased by %10 in the last year (target class) whose sales decreased by at least %30 during the same period (contrasting class) Example 2: Compare two groups of AE customers I) who shop for computer products regularly (target class) more than two times a month II) who rarely shop for such products (contrasting class) less than three times a year The resulting description: %80 of I group customers university education ages 20-40 %60 of II group customers seniors or young no university degree

Multidimensional Data
sales according to region month and product type Dimensions: Product, Location, Time Hierarchical summarization paths Region Industry Region Year Category Country Quarter Product City Month Week Office Day Product Month

Association Analysis Discovery of association rules showing attribute-value conditions that occur frequently together in a given set of data widely used market basket transaction data analysis more formally X  Y that is A1A2.. Ak  B1B2.. Bl A1 , B1 are attribute value pairs or predicates

Example: association analysis
From the AllEs database age(X,”20..29”)income(X,”1, ,000”)buy(X,”Notebook computer”) (support = %2, confidence= %60) X is a variable representing a customer %2 of the AE customers are between 20 and 29 age incomes ranging from 1 to 2 billon TL Buy Notebook with %60 probability that customers in those age and income groups will buynote book a multidimensional association rule contains more than one attribute or predicate

Market basket analysis
customers buying behaviour is investigated Based on only the transactions data no information about customer properties: age income Managers are interested in which products or product groups are sold together

Transactional Database
Transaction ID Item List 10001 Computer,NB,pritner 10002 Ploter,monitor,mouse 10003 Computer,NB 10004 Printer 10005 Ploter,UPS,modem

Example: basket analysis rule
buy(notebok)buy(printer) (support= %1,confidence=%60) %1 of all transactions contains computer and printer if a transaction containsnotebook there is a %60 chance that it contains printer as well a single dimensional association rule contains a single predicate an association rule is interesting if its support exceeds a minimum threshold and its confidence exceeds a min threshold These min values are set by specialists

Classification Learning is supervised
Dependent variable is categorical Build a model able to assign new instances to one of a set of well-defined classes

Typical Classification Problems
Given characteristics of individuals differentiate them who have suffered a heart attack from those who have not Determine if a credit card purchase is fraudulent Classify a car loan applicant as a good or a poor credit risk

Methods of Classification
Decision Trees Artificial Neural Networks Bayesian Classification Naïve Belief Networks k-nearest neighbor Regression Logistic (logit) probit Predicts probability of each class when the dependent variable is categorical good customer bed customer or employed unemployed

Steps of classification process
(1) Train the model using a training set data objects whose class labels are known (2) Test the model on a test sample whose class labels are known but not used for training the model (3) Use the model for classification on new data whose class labels are unknown

An example - classification
Cust ID age income education Type 1 35 800 udergrad risky 2 26 600 HighSch 3 48 1200 grad normal 8 52 2500 good 44 29 1700 Historical data Each customer type İs known Each customer has a Label CustID age income education Type 17 43 550 Ph.D. risky 27 68 1650 grad Normal Testing set whose labels are also Known but not used in model Training the model CustID age income Educatin Type 11 36 850 Udergrd ? 27 28 1650 grad New customers Whose type hsa to be Estimated Each new customer hss to be classified as Risky normal or good

Orginal data

Historical data Each customer type İs known
Each customer has a Label Testing set whose labels are also Known but not used in model Training the model New customers Whose type hsa to be Estimated Each new customer hss to be classified as buyer or non buyer

An example – classification cont.
Based on historical data develop a classification model Decision tree, neural network, regression ... Test the performance of the model on a portion of the historical data İf accuricy of the model is satisfactory Use the model on the new customers 11 and 27 to assign a type the these new customers

Example AE customers age goodl risky Yearly income

Example AE customers age goodl risky Yearly income ?
Assign the new customer whose type in unknown to either * or +

Solution x2 : age x1 : yearly income 1000 good risky 35 rule: IF yearly income> 1000 and age> 35 THEN good ELSE risky

Credit Card Promotion Policy
Credit card companies Promotional offerings with their monthly credit card billing Offers provide the opportunity to purchase items such as magazines, … A data mining study Predict individual behaviour What is the likelihood of an individual towards taking the advantage of promotions based on individual characteristics, credit history.. Expected reduction in postage; paper and processing costs for the credit card company

Lıfe Insurance Promotıon
Credit Card Promotion Database Income Range Magazıne Promotıon Watch Promotıon Lıfe Insurance Promotıon Gender Age Credıt Card Insurance 40-50 K Yes No Male 45 30-40 K Female 40 42 43 50-60 K 38 20-30 K 55 35 27 41 29 39 19

Decision Trees for Credit Card Insurance Database
age Dependent Variable Life Insurance Promotion <=43 >43 critical value of 43 is deter by the algorithm N 3,Y 0 Decision:No Gender Female Male A Production Rule from the Tree IF (age<=43)&(Sex=Male) &(Credit Card In = No) THEN Life Insurance Pr = No Cr Ins N 0, Y 6 Decision: Yes Yes No Yes 2, No 0 Decision? Yes N 4, Y 1 Decision: No

Artificial Neural Networks
Set of interconnected nodes designed to imitate the functioning of the human brain Feed-forward network Supervised learner model

For the promotion example
Encode all variables Assign a numerical value even for qualitative variables such as sex Say X1 represent gender When Male X1 =1 Female X1 =0

Input layer Hidden layer Output layer 1 W1,5=0.014 X1=+1 5 W5,9=-0.17 X2=0 X3=0.5 X4=-1 (1-0.78)2 is error square 1 actual value of O9 for a particular Data object 0.78 is predicted value

Weights updating Weights between nodes are adjusted so as to reduce error Details of the training process for neural networks are not important for the time being

Estimation-Prediction
Similar to classification Output is a continuous variable Estimation: current value Prediction: future outcome rather then current behavior

Typical Estimation-Prediction Problems
Estimate the salary of an individual who owns a sports car Predict next week`s closing price for the IMKB100 index Forecast next days temperature

Prediction methods Artificial Neural networks linear regression
Yi = a0+a1X1,i+a2X2,i+...+akXk,i+ui non-linear regression Yi =f(X1,i, X2,i,.., Xk,ia1,a2,..,ak,ui) generalized linear regression logistic logit,probit poisson regression for count variables Regression Trees

Example:Prediction and Classification
Classification is used to classify customers applying for credit cards known class labels: risky,reliable when a new customer applies looking at her charecteristics income age education wealth region ... Customer class is predicted Prediction: The monthly expense of a new customer ( a real continuous variable ) is predicted based on personal information independent variables income education wealth profession ... Some are numeric some categorical

Cluster Analysis Class label is unknown: Group data to form new classes, assign class labels to each data object Unknown generated by the clustering model e.g., cluster customers to find customer segments Clustering based on the principle: maximizing the intra-class similarity and minimizing the interclass similarity Objects within a cluster have high similarity in comparison to one another but are very dissimilar to objects in other clusters there may be hierarchy of classes

Example: Clustering Can be performed on AE customer data
to identify homogenous subpopulations of customers represent individual target groups for marketing

Before clustering After clustering

Clustering according to income and distance to store
Type1 Type 2 type 3 income Clustering according to income and distance to store three cluster of data points are evident

Outlier Analysis Outlier: a data object that does not comply with the general behavior of the data It can be considered as noise or exception but is quite useful in fraud detection, rare events analysis DECTECED using statistical tests distance measures visually inspecting the data Examples:

Reasons for outliers Measurement errors coding errors nature of data
age is entered as 999 nature of data salary of the general manager is much more higher than the other employees in crisis the interest rate was in the order of 1000s

Evolution Analysis Describes and models regularities or trends for objects whose behavior changes over time Distinct features include Trend and deviation: time-series data analysis Sequential pattern mining, periodicity analysis Similarity-based analysis Example Stock market predictions: future stock prices for overall stocks: indexes or individual company stocks

Sequential Pattern Analysis
Determine sequential patterns in data Based on time sequence of actions Similar to associations Relationship is based on time Example 1: buy CD player today buy CD within one week Example 2: In what sequence web pages of an e-business company are accessed %70 percents of visitors follows A B C or A D B C or A E B C He then determines to add a link directly from page A to page C

Are All the “Discovered” Patterns Interesting?
A data mining system/query may generate thousands of patterns, not all of them are interesting. Are all patterns interesting? Typically not -only a small fraction of patterns are interesting to any given user Interestingness measures: A pattern is interesting if it is easily understood by humans, valid on new or test data with some degree of certainty, potentially useful, novel, or validates some hypothesis that a user seeks to confirm

Objective vs. subjective interestingness measures:
Objective: based on statistics and structures of patterns, e.g., support, X Y P(X  Y):probability of a transaction contains both X and Y confidence, degree of certainty of the detected association P(Y I X) the conditional probability : the probability that a transaction containing X also contains Y thresholds - controlled by the user ex: rules that do not satisfy a confidence threshold of %50 are uninteresting Subjective: based on user’s belief in the data, e.g., unexpectedness, novelty, actionability, etc.

Methodology of Knowledge Discovery in Databases Data mining functionalities Are all the patterns interesting? Business Applications of data mining

Potential Business Applications
Market analysis and management target marketing, customer relation management, market basket analysis, cross selling, market segmentation Risk analysis and management Banks assume a financial risk when they grant loans risk models attempt to predict the probability of default or fail to pay back the borrowed amount Credit cards Insurance companies Fraud detection and management Other Applications Text mining (news group, , documents) and Web analysis. Intelligent query answering

Market Analysis and Management (1)
Where are the data sources for analysis? Credit card transactions, loyalty cards, discount coupons, customer complaint calls, plus (public) lifestyle studies,clickstreams Customer profiling-segmentation data mining can tell you what types of customers buy what products (clustering or classification) Target marketing Find clusters of “model” customers who share the same characteristics: interest, income level, spending habits, etc.

Effectiveness of sales campaigns Advertisements, coupons, discounts, bonuses promote products and attract customers can help improve profits Compare amount of sales and number of transactions during the sales period versus before or after the sales campaign Association analysis which items are likely to be purchased together with the items on sale

Customer retention Analysis of Customer loyalty sequences of purchases of particular customers goods purchased at different periods by the same customers can be grouped into sequences changes in customer consumption or loyalty suggests adjustments on the pricing and variety of goods to retain old customers and attract new customers Cross-selling and up-selling associations from sales records a customer who buy a PC is likely to buy a printer purchase recommendations

Fraud Detection and Management
Applications widely used in health care, retail, credit card services, telecommunications (phone card fraud), etc. Approach use historical data to build models of fraudulent behavior and use data mining to help identify similar instances Examples Credit card transactions: The FALCON fraud assessment system by HNC Inc. to signal possibly fraudulent credit card transactions money laundering: detect suspicious money transactions (US Treasury's Financial Crimes Enforcement Network) Detecting telephone fraud:ASPECT European Research Gr. Unsupervised clustering to detect fraud in mobile phone networks Telephone call model: destination of the call, duration, time of day or week. Analyze patterns that deviate from an expected norm.

Health Care Storing patients` records in electronic format, developments in medical information systems Large amount of clinical data Regularities, trends and surprising events extracted by data mining methods ANN, temporal reasoning assist clinicians to make informed decisions and improving health sevices MERCK-MEDCO Managed Care, Pharmaceutical Insurance … company Uncover less expensive but equally effective drug treatments

Financial Data Analysis
complete, reliable, high quality Loan payment prediction and customer credit policy analysis

Loan payment prediction and customer credit policy analysis
Factors influencing loan payment performance loan-to-value ratio term of the loan dept ratio (total monthly debt/total monthly income) payment-to-income ratio income level education level residence region credit history analysis may find that payment-income ratio is a dominant factor while education level and debt ratio are not

Risk Management and Insurance
determine insurance rates manage investment portfolios differentiate between companies and/or individuals who are good and poor credit risks Farmer`s Group discover a scenario: Someone who owns a sports car is not a higher accident risk Conditions: the sport car to be a second car and the family car to be a station wagon or a sedan

Data Mining for the Telecommunication Industry
Telecommunication data are multidimensional calling-time duration location of caller location of callee type of call used to identify and compare data traffic system workload resource usage user group behavior profit fraudulent pattern analysis and identification of unusual patterns to achieve customer loyalty characteristics of customers affecting line usage

Other Applications Sports and Gaming Text Mining Internet Web Mining
Predicting outcome of football games Text Mining Spam detection Internet Web Mining Web usage mining İmprove link structure Recommander Systmes Web structure mining: mining link structure of Web

Other Applications Educational Data Mining Human Resources
Clustering students Design enterece exams, selection policies Human Resources How to select applicants Online Dating Recommandataions to visitors

Summary Data mining: discovering interesting patterns from large amounts of data A natural evolution of database technology, in great demand, with wide applications A KDD process includes data cleaning, data integration, data selection, transformation, data mining, pattern evaluation, and knowledge presentation Mining can be performed in a variety of information repositories Data mining functionalities: characterization, discrimination, association, classification, clustering, outlier and trend analysis, etc. Classification of data mining systems Major issues in data mining

Chapter I:Introduction MIS /2017 Fall

Similar presentations

Presentation on theme: "Chapter I:Introduction MIS /2017 Fall"— Presentation transcript:

Similar presentations

About project

Feedback

Log in

Auth with social network:

Chapter I:Introduction MIS /2017 Fall

Similar presentations

Presentation on theme: "Chapter I:Introduction MIS /2017 Fall"— Presentation transcript:

Similar presentations

About project

Feedback