Topics: What is concept description? Data generalization and summarization-basedcharacterization Analytical characterization: Analysis of attribute relevance Mining class comparisons: Discriminating betweendifferent classes Mining descriptive statistical measures in large databases Discussion SummaryWhat is Concept Description? Descriptive vs. predictive data mining Descriptive mining: describes concepts or task-relevantdata sets in concise, summarative, informative,discriminative forms Predictive mining: Based on data and analysis,constructs models for the database, and predicts thetrend and properties of unknown data Concept description: Characterization: provides a concise and succinctsummarization of the given collection of data Comparison: provides descriptions comparing two ormore collections of data4Concept Description vs. OLAP Concept description: can handle complex data types of theattributes and their aggregations a more automated process OLAP: restricted to a small number of dimension andmeasure types user-controlled process5DATA GENERALIZATION ANDSUMMARIZATION-BASEDCHARACTERIZATION6Data Generalization and SummarizationbasedCharacterization Data generalization A process which abstracts a large set of task-relevantdata in a database from a low conceptual levels tohigher ones. Approaches: Data cube approach(OLAP approach) Attribute-oriented induction approach12345Conceptual levels7Characterization: Data Cube Approach Data are stored in data cube Identify expensive computations e.g., count( ), sum( ), average( ), max( ) Perform computations and store results in datacubes Generalization and specialization can beperformed on a data cube by roll-up and drilldown An efficient implementation of datageneralization8Data Cube Approach (Cont…) Limitations can only handle data types of dimensions tosimple nonnumeric data and of measures tosimple aggregated numeric values . Lack of intelligent analysis, can’t tell whichdimensions should be used and what levelsshould the generalization reach9Attribute-Oriented Induction Proposed in 1989 (KDD ‘89 workshop) Not confined to categorical data nor particular measures. How it is done? Collect the task-relevant data (initial relation) using arelational database query Perform generalization by attribute removal orattribute generalization. Apply aggregation by merging identical, generalizedtuples and accumulating their respective counts Interactive presentation with usersBasic Principles of Attribute-Oriented Induction Data focusing: task-relevant data, including dimensions, andthe result is the initial relation. Attribute-removal: remove attribute A if there is a large setof distinct values for A but (1) there is no generalizationoperator on A, or (2) A’s higher level concepts areexpressed in terms of other attributes. Attribute-generalization: If there is a large set of distinctvalues for A, and there exists a set of generalizationoperators on A, then select an operator and generalize A. Attribute-threshold control: typical 2-8, specified/default. Generalized relation threshold control: control the finalrelation/rule size. see exampleAttribute-Oriented Induction: Basic Algorithm InitialRel: Query processing of task-relevant data, derivingthe initial relation. PreGen: Based on the analysis of the number of distinctvalues in each attribute, determine generalization plan foreach attribute: removal? or how high to generalize? PrimeGen: Based on the PreGen plan, performgeneralization to the right level to derive a “primegeneralized relation”, accumulating the counts. Presentation: User interaction: (1) adjust levels by drilling,(2) pivoting, (3) mapping into rules, cross tabs,visualization presentations.12Example DMQL: Describe general characteristics of graduatestudents in the Big-University databaseuse Big_University_DBmine characteristics as “Science_Students”in relevance to name, gender, major, birth_place,birth_date, residence, phone#, gpafrom studentwhere status in “graduate” Corresponding SQL statement:Select name, gender, major, birth_place, birth_date,residence, phone#, gpafrom studentwhere status in “Msc”, “MBA”, “PhD” Class Characterization: An ExampleName Gender Major Birth-Place Birth_date Residence Phone # GPAJimWoodmanM CS Vancouver,BC,Canada8-12-76 3511 Main St.,Richmond687-4598 3.67ScottLachanceM CS Montreal, Que,Canada28-7-75 345 1st Ave.,Richmond253-9106 3.70Laura Lee…F…Physics…Seattle, WA, USA…25-8-70…125 Austin Ave.,Burnaby…420-5232…3.83…Removed Retained Sci,Eng,BusCountry Age range City Removed Excl,VG,..Gender Major Birth_region Age_range Residence GPA CountM Science Canada 20-25 Richmond Very-good 16F Science Foreign 25-30 Burnaby Excellent 22… … … … … … …Birth_RegionGenderCanada Foreign TotalM 16 14 30F 10 22 32Total 26 36 62PrimeGeneralizedRelationInitialRelationPresentation of Generalized Results Generalized relation: Relations where some or all attributes are generalized, with countsor other aggregation values accumulated. Cross tabulation: Mapping results into cross tabulation form (similar to contingencytables). Visualization techniques: Pie charts, bar charts, curves, cubes, and other visual forms. Quantitative characteristic rules: Mapping generalized result into characteristic rules with quantitativeinformation associated with it, e.g.,_ ( ) ” “[ :53%] _ ( ) ” “[ :47%].( ) ( )birth region x Canada t birth region x foreign tgrad x male x 15Presentation—Generalized Relation16Presentation—Crosstab17Implementation by Cube Technology Construct a data cube on-the-fly for the given datamining query Facilitate efficient drill-down analysis May increase the response time A balanced solution: precomputation of “subprime”relation Use a predefined & precomputed data cube Construct a data cube beforehand Facilitate not only the attribute-oriented induction, butalso attribute relevance analysis, dicing, slicing, rollupand drill-down Cost of cube computation and the nontrivial storageoverhead18ANALYTICAL CHARACTERIZATION:ANALYSIS OF ATTRIBUTERELEVANCE19Characterization vs. OLAP Similarity: Presentation of data summarization at multiple levels ofabstraction. Interactive drilling, pivoting, slicing and dicing. Differences: Automated desired level allocation. Dimension relevance analysis and ranking when thereare many relevant dimensions. Sophisticated typing on dimensions and measures. Analytical characterization: data dispersion analysis.20Attribute Relevance Analysis Why? Which dimensions should be included? How high level of generalization? Automatic VS. Interactive Reduce # attributes; Easy to understand patterns What? statistical method for preprocessing data filter out irrelevant or weakly relevant attributes retain or rank the relevant attributes relevance related to dimensions and levels analytical characterization, analytical comparison21Attribute relevance analysis (cont’d) How? Data Collection Analytical Generalization Use information gain analysis (e.g., entropy or othermeasures) to identify highly relevant dimensions and levels. Relevance Analysis Sort and select the most relevant dimensions and levels. Attribute-oriented Induction for class description On selected dimension/level OLAP operations (e.g. drilling, slicing) on relevancerules22Relevance Measures Quantitative relevance measure determines theclassifying power of an attribute within a set ofdata. Methods information gain (ID3) gain ratio (C4.5) gini index 2 contingency table statistics uncertainty coefficient23Information-Theoretic Approach Decision tree each internal node tests an attribute each branch corresponds to attribute value each leaf node assigns a classification ID3 algorithm build decision tree based on training objectswith known class labels to classify testingobjects rank attributes with information gain measure minimal height the least number of tests to classify an object24Top-Down Induction of Decision TreeAttributes = Outlook, Temperature, Humidity, WindOutlookHumidity Windsunny overcast rainyesno yeshigh normalnostrong weakyesPlayTennis = yes, no25Entropy and Information Gain S contains si tuples of class Ci for i = 1, …, m Information measures info required to classifyany arbitrary tuple Entropy of attribute A with values a1,a2,…,av Information gained by branching on attribute Aslog ssI( s ,s ,…,s ) s m iii1 2 m 21 I( s ,…,s )sE(A) s … s j mjvjj mj111 Gain(A) I(s1,s 2,…,sm) E(A)26Example: Analytical Characterization Task Mine general characteristics describing graduatestudents using analytical characterization Given attributes name, gender, major, birth_place, birth_date,phone#, and gpa Gen(ai) = concept hierarchies on ai Ui = attribute analytical thresholds for ai Ti = attribute generalization thresholds for ai R = attribute relevance threshold27Example: Analytical Characterization (cont’d) 1. Data collection target class: graduate student contrasting class: undergraduate student 2. Analytical generalization using Ui attribute removal remove name and phone# attribute generalization generalize major, birth_place, birth_date and gpa accumulate counts candidate relation: gender, major, birth_country,age_range and gpa28Example: Analytical characterization (2)gender major birth_country age_range gpa countM Science Canada 20-25 Very_good 16F Science Foreign 25-30 Excellent 22M Engineering Foreign 25-30 Excellent 18F Science Foreign 25-30 Excellent 25M Science Canada 20-25 Excellent 21F Engineering Canada 20-25 Excellent 18Candidate relation for Target class: Graduate students (=120)gender major birth_country age_range gpa countM Science Foreign <20 Very_good 18 F Business Canada <20 Fair 20 M Business Canada <20 Fair 22 F Science Canada 20-25 Fair 24 M Engineering Foreign 20-25 Very_good 22 F Engineering Canada <20 Excellent 24 Candidate relation for Contrasting class: Undergraduate students (=130) 29 Example: Analytical characterization (3) 3. Relevance analysis Calculate expected info required to classify an arbitrary tuple Calculate entropy of each attribute: e.g. major 0 9988 250 130 250 130 250 120 250 120130 120 I(s1,s 2 ) I( , ) log 2 log 2 . For major=”Science”: S11=84 S21=42 I(s11,s21)=0.9183 For major=”Engineering”: S12=36 S22=46 I(s12,s22)=0.9892 For major=”Business”: S13=0 S23=42 I(s13,s23)=0 Number of grad students in “Science” Number of undergrad students in “Science” 30 Example: Analytical Characterization (4) Calculate expected info required to classify a given sample if S is partitioned according to the attribute Calculate information gain for each attribute Information gain for all attributes 0 7873 250 42 250 82 250 126 E(major) I( s11,s21 ) I( s12,s22 ) I( s13,s23 ) . Gain(major) I(s1,s 2 ) E(major) 0.2115 Gain(gender) = 0.0003 Gain(birth_country) = 0.0407 Gain(major) = 0.2115 Gain(gpa) = 0.4490 Gain(age_range) = 0.5971 31 Example: Analytical characterization (5) 4. Initial working relation (W0) derivation R = 0.1 remove irrelevant/weakly relevant attributes from candidate relation => drop gender, birth_country remove contrasting class candidate relation 5. Perform attribute-oriented induction on W0 using Timajor age_range gpa countScience 20-25 Very_good 16Science 25-30 Excellent 47Science 20-25 Excellent 21Engineering 20-25 Excellent 18Engineering 25-30 Excellent 18Initial target class working relation W0: Graduate students32MINING CLASS COMPARISONS:DISCRIMINATING BETWEENDIFFERENT CLASSESMining Class Comparisons Comparison: Comparing two or more classes Method: Partition the set of relevant data into the target class and thecontrasting class(es) Generalize both classes to the same high level concepts Compare tuples with the same high level descriptions Present for every tuple its description and two measures support – distribution within single class comparison – distribution between classes Highlight the tuples with strong discriminant features Relevance Analysis: Find attributes (features) which best distinguish different classes34Example: Analytical comparison Task Compare graduate and undergraduate students usingdiscriminant rule. DMQL queryuse Big_University_DBmine comparison as “grad_vs_undergrad_students”in relevance to name, gender, major, birth_place, birth_date, residence, phone#, gpafor “graduate_students”where status in “graduate”versus “undergraduate_students”where status in “undergraduate”analyze count%from student35Example: Analytical comparison (2) Given attributes name, gender, major, birth_place,birth_date, residence, phone# and gpa Gen(ai) = concept hierarchies on attributes ai Ui = attribute analytical thresholds forattributes ai Ti = attribute generalization thresholds forattributes ai R = attribute relevance threshold36Example: Analytical comparison (3) 1. Data collection target and contrasting classes 2. Attribute relevance analysis remove attributes name, gender, major, phone# 3. Synchronous generalization controlled by user-specified dimension thresholds prime target and contrasting class(es)relations/cuboids37Example: Analytical comparison (4)Birth_country Age_range Gpa Count%Canada 20-25 Good 5.53%Canada 25-30 Good 2.32%Canada Over_30 Very_good 5.86%… … … …Other Over_30 Excellent 4.68%Prime generalized relation for the target class: Graduate studentsBirth_country Age_range Gpa Count%Canada 15-20 Fair 5.53%Canada 15-20 Good 4.53%… … … …Canada 25-30 Good 5.02%… … … …Other Over_30 Excellent 0.68%Prime generalized relation for the contrasting class: Undergraduate students38Example: Analytical comparison (5) 4. Drill down, roll up and other OLAP operationson target and contrasting classes to adjust levelsof abstractions of resulting description 5. Presentation as generalized relations, crosstabs, bar charts,pie charts, or rules contrasting measures to reflect comparisonbetween target and contrasting classes e.g. count%39Quantitative Discriminant Rules Cj = target class qa = a generalized tuple covers some tuples of class but can also cover some tuples of contrasting class d-weight range: [0, 1] quantitative discriminant rule form mia ia jcount(q C )d weight count(q C )1 X, target_class(X)condition(X) [d : d_weight]40Example: Quantitative Discriminant Rule Quantitative discriminant rule where 90/(90+210) = 30%Status Birth_country Age_range Gpa CountGraduate Canada 25-30 Good 90Undergraduate Canada 25-30 Good 210Count distribution between graduate and undergraduate students for a generalized tuple_ ( ) ” ” _ ( ) “25 30″ ( ) ” ” [ : 30%], _ ( )birth country X Canada age range X gpa X good dX graduate student X 41Class Description Quantitative characteristic rule necessary Quantitative discriminant rule sufficient Quantitative description rule necessary and sufficient1 [t :w1,d :w1]… [t :wn,d :wn] condition (X) condition (X)X, target_class(X)n X, target_class(X)condition(X) [d : d_weight] X, target_class(X)condition(X) [t : t_weight]42Example: Quantitative Description Rule Quantitative description rule for target class EuropeLocation/item TV Computer Both_itemsCount t-wt d-wt Count t-wt d-wt Count t-wt d-wtEurope 80 25% 40% 240 75% 30% 320 100% 32%N_Am 120 17.65% 60% 560 82.35% 70% 680 100% 68%Both_regions200 20% 100% 800 80% 100% 1000 100% 100%Crosstab showing associated t-weight, d-weight values and total number(in thousands) of TVs and computers sold at AllElectronics in 1998(item(X) “TV” )t : 25%,d : 40%[t : 75%,d : 30%]X,Europe(X) 43Mining Complex Data Objects:Generalization of Structured Data Set-valued attribute Generalization of each value in the set into its correspondinghigher-level concepts Derivation of the general behavior of the set, such as thenumber of elements in the set, the types or value ranges inthe set, or the weighted average for numerical data E.g., hobby = tennis, hockey, chess, violin, nintendo_games generalizes to sports, music, video_games List-valued or a sequence-valued attribute Same as set-valued attributes except that the order of theelements in the sequence should be observed in thegeneralization44Generalizing Spatial and Multimedia Data Spatial data: Generalize detailed geographic points into clustered regions,such as business, residential, industrial, or agricultural areas,according to land usage Require the merge of a set of geographic areas by spatialoperations Image data: Extracted by aggregation and/or approximation Size, color, shape, texture, orientation, and relative positionsand structures of the contained objects or regions in the image Music data: Summarize its melody: based on the approximate patterns thatrepeatedly occur in the segment Summarized its style: based on its tone, tempo, or the majormusical instruments played45Generalizing Object Data Object identifier: generalize to the lowest level of class in theclass/subclass hierarchies Class composition hierarchies generalize nested structured data generalize only objects closely related in semantics to the currentone Construction and mining of object cubes Extend the attribute-oriented induction method Apply a sequence of class-based generalization operators ondifferent attributes Continue until getting a small number of generalized objects thatcan be summarized as a concise in high-level terms For efficient implementation Examine each attribute, generalize it to simple-valued data Construct a multidimensional data cube (object cube) Problem: it is not always desirable to generalize a set of values tosingle-valued data46An Example: Plan Mining by Divide & Conquer Plan: a variable sequence of actions E.g., Travel (flight): Plan mining: extraction of important or significant generalized(sequential) patterns from a planbase (a large collection of plans) E.g., Discover travel patterns in an air flight database, or find significant patterns from the sequences of actions in the repairof automobiles Method Attribute-oriented induction on sequence data A generalized travel plan: <small-big-small> Divide & conquer:Mine characteristics for each subsequence E.g., big: same airline, small-big: nearby region47A Travel Database for Plan Mining Example: Mining a travel planbaseplan# action# departure depart_time arrival arrival_time airline …1 1 ALB 800 JFK 900 TWA …1 2 JFK 1000 ORD 1230 UA …1 3 ORD 1300 LAX 1600 UA …1 4 LAX 1710 SAN 1800 DAL …2 1 SPI 900 ORD 950 AA …. . . . . . . .. . . . . . . .. . . . . . . .airport_code city state region airport_size …1 1 ALB 800 …1 2 JFK 1000 …1 3 ORD 1300 …1 4 LAX 1710 …2 1 SPI 900 …. . . . .. . . . .. . . . .Travel plans tableAirport info table48Multidimensional Analysis Strategy Generalize theplanbase indifferentdirections Look forsequentialpatterns in thegeneralized plans Derive high-levelplansA multi-D model for the planbase49Multidimensional GeneralizationPlan# Loc_Seq Size_Seq State_Seq1 ALB – JFK – ORD – LAX – SAN S – L – L – L – S N – N – I – C – C2 SPI – ORD – JFK – SYR S – L – L – S I – I – N – N. . .. . .. . .Multi-D generalization of the planbasePlan# Size_Seq State_Seq Region_Seq …1 S – L+ – S N+ – I – C+ E+ – M – P+ …2 S – L+ – S I+ – N+ M+ – E+ …. . .. . .. . .Merging consecutive, identical actions in plans( ) ( ) [75%]( , ,) _ ( , ) _ ( , )region x region yflight x y airport size x S airport size y L 50Generalization-Based Sequence Mining Generalize planbase in multidimensional way usingdimension tables Use # of distinct values (cardinality) at each level todetermine the right level of generalization(level-“planning”) Use operators merge “+”, option “[]” to further generalizepatterns Retain patterns with significant support51Generalized Sequence Patterns AirportSize-sequence survives the min threshold (afterapplying merge operator):S-L+-S [35%], L+-S [30%], S-L+ [24.5%], L+ [9%] After applying option operator:[S]-L+-[S] [98.5%] Most of the time, people fly via large airports to get tofinal destination Other plans: 1.5% of chances, there are other patterns:S-S, L-S-L52MINING DESCRIPTIVESTATISTICAL MEASURES IN LARGEDATABASES53Mining Data Dispersion Characteristics Motivation To better understand the data: central tendency, variationand spread Data dispersion characteristics median, max, min, quantiles, outliers, variance, etc. Numerical dimensions correspond to sorted intervals Data dispersion: analyzed with multiple granularities ofprecision Boxplot or quantile analysis on sorted intervals Dispersion analysis on computed measures Folding measures into numerical dimensions Boxplot or quantile analysis on the transformed cube54Measuring the Central Tendency Mean Weighted arithmetic mean Median: A holistic measure Middle value if odd number of values, or average of themiddle two values otherwise estimated by interpolation Mode Value that occurs most frequently in the data Unimodal, bimodal, trimodal Empirical formula:nii xnx11 niinii iww xx11cfn f lmedian Lmedian)/ 2 ( )( 1 mean mode 3(mean median)55Measuring the Dispersion of Data Quartiles, outliers and boxplots Quartiles: Q1 (25th percentile), Q3 (75th percentile) Inter-quartile range: IQR = Q3 – Q1 Five number summary: min, Q1, M, Q3, max Boxplot: ends of the box are the quartiles, median is marked,whiskers, and plot outlier individually Outlier: usually, a value higher/lower than 1.5 x IQR Variance and standard deviation Variance s2: (algebraic, scalable computation) Standard deviation s is the square root of variance s2 ninii inii xnxnx xns1 12 212 2 [ 1 ( ) ]1( ) 11156Boxplot Analysis Five-number summary of a distribution:Minimum, Q1, M, Q3, Maximum Boxplot Data is represented with a box The ends of the box are at the first and thirdquartiles, i.e., the height of the box is IRQ The median is marked by a line within the box Whiskers: two lines outside the box extend toMinimum and Maximum57Visualization of Data Dispersion: Boxplot Analysis58Mining Descriptive Statistical Measures in LargeDatabases Variance Standard deviation: the square root of the variance Measures spread about the mean It is zero if and only if all the values are equal Both the deviation and the variance are algebraic 2 212 2 11( ) 111i inii xnxnx xns59Histogram Analysis Graph displays of basic statistical class descriptions Frequency histograms A univariate graphical method Consists of a set of rectangles that reflect the counts orfrequencies of the classes present in the given data60Quantile Plot Displays all of the data (allowing the user to assess boththe overall behavior and unusual occurrences) Plots quantile information For a data xi data sorted in increasing order, fiindicates that approximately 100 fi% of the data arebelow or equal to the value xi61Quantile-Quantile (Q-Q) Plot Graphs the quantiles of one univariate distribution againstthe corresponding quantiles of another Allows the user to view whether there is a shift in goingfrom one distribution to another62Scatter plot Provides a first look at bivariate data to see clusters ofpoints, outliers, etc Each pair of values is treated as a pair of coordinates andplotted as points in the plane63Loess Curve Adds a smooth curve to a scatter plot in order toprovide better perception of the pattern of dependence Loess curve is fitted by setting two parameters: asmoothing parameter, and the degree of thepolynomials that are fitted by the regression64Graphic Displays of Basic Statistical Descriptions Histogram: (shown before) Boxplot: (covered before) Quantile plot: each value xi is paired with fi indicatingthat approximately 100 fi % of data are xi Quantile-quantile (q-q) plot: graphs the quantiles of oneunivariant distribution against the corresponding quantilesof another Scatter plot: each pair of values is a pair of coordinatesand plotted as points in the plane Loess (local regression) curve: add a smooth curve to ascatter plot to provide better perception of the pattern ofdependence65DISCUSSION66AO(Attribute Oriented) Induction vs.Learning-from-example Paradigm Difference in philosophies and basic assumptions Positive and negative samples in learning-fromexample:positive used for generalization, negative –for specialization Positive samples only in data mining: hencegeneralization-based, to drill-down backtrack thegeneralization to a previous state Difference in methods of generalizations Machine learning generalizes on a tuple by tuple basis Data mining generalizes on an attribute by attributebasis67Entire vs. Factored Version Space68Incremental and Parallel Mining of ConceptDescription Incremental mining: revision based on newly added dataDB Generalize DB to the same level of abstraction in thegeneralized relation R to derive R Union R U R, i.e., merge counts and other statisticalinformation to produce a new relation R’ Similar philosophy can be applied to data sampling,parallel and/or distributed mining, etc.69Summary Concept description: characterization and discrimination OLAP-based vs. attribute-oriented induction Efficient implementation of AOI Analytical characterization and comparison Mining descriptive statistical measures in largedatabases Discussion Incremental and parallel mining of description Descriptive mining of complex types of data70Thank you !!!Questions71
Explain analytical characterization?Methods of attribute relevance analysis?How does analytical datacharacterization/comparison performs?From the descriptive statistics point of view, why is itthat additional statistical measures should beintroduced in describing central tendency and datadispersion? Give an example.In comparison with machine learning algorithm, whyis it that database-oriented concept description leadsto efficiency and scalability in large databases anddata warehouse?72Discuss why analytical data characterization is needed andhow it can be performed. Compare the result of twoinduction methods with relevance analysis and withoutrelevance analysis.Give three additional commonly used statistical measuresfor the characterization of data dispersion and discuss howthey can be computed efficiently in large databases.[Button id=”1″]
Quality and affordable writing services. Our papers are written to meet your needs, in a personalized manner. You can order essays, annotated bibliography, discussion, research papers, reaction paper, article critique, coursework, projects, case study, term papers, movie review, research proposal, capstone project, speech/presentation, book report/review, and more.
Need Help? Click On The Order Now Button For Help
What Students Are Saying About Us
.......... Customer ID: 12*** | Rating: ⭐⭐⭐⭐⭐"Honestly, I was afraid to send my paper to you, but splendidwritings.com proved they are a trustworthy service. My essay was done in less than a day, and I received a brilliant piece. I didn’t even believe it was my essay at first 🙂 Great job, thank you!"
.......... Customer ID: 14***| Rating: ⭐⭐⭐⭐⭐
"The company has some nice prices and good content. I ordered a term paper here and got a very good one. I'll keep ordering from this website."
