0% found this document useful (0 votes)
9 views17 pages

Data Preprocessing Techniques Explained

The document provides an overview of data preprocessing in data mining, detailing its importance and steps including data cleaning, integration, reduction, and transformation. It emphasizes data quality measures such as accuracy, completeness, and consistency, and discusses methods for handling missing and noisy data. Additionally, it covers strategies for data reduction and transformation techniques like normalization and attribute construction.

Uploaded by

mehrmehsaqib
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
9 views17 pages

Data Preprocessing Techniques Explained

The document provides an overview of data preprocessing in data mining, detailing its importance and steps including data cleaning, integration, reduction, and transformation. It emphasizes data quality measures such as accuracy, completeness, and consistency, and discusses methods for handling missing and noisy data. Additionally, it covers strategies for data reduction and transformation techniques like normalization and attribute construction.

Uploaded by

mehrmehsaqib
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Data Mining

DATA PREPROCESSING

Department of Information Technology


UNIVERSI T Y OF SARGODAHA
Lesson from Holy Quran
2
Outline
3

 Introduction to Data Preprocessing


 Data Quality
 Steps of Data Preprocessing
 Data cleaning
 Data integration
 Data reduction
 Data transformation
4 Data Preprocessing
Data Quality
5

 Measures for data quality:


 Accuracy: correct or wrong
 Completeness: not recorded, unavailable, …
 Consistency: some updated/modified but some not.
 Timeliness: timely update?
 Believability: how trustable the data is?
 Interpretability: how easily data can be understood?
Major Tasks in Data Preprocessing
6

 Data cleaning
 Fill in missing values, smooth noisy data, identify or remove outliers, and resolve
inconsistencies
 Data integration
 Integration of multiple databases or files, diverse sources
 Data reduction
 Dimensionality reduction
 Data compression
 Data transformation
 Normalization
Data Cleaning
7

 Data in the Real World Is Dirty: (More thanks to Social Web)


 Lots of potentially incorrect data, e.g., human or computer error, extraction error
 incomplete: lacking attribute values,
◼ e.g., Occupation=“ ” (missing data)
 noisy: containing noise, errors
◼ e.g., Salary=“−10” (an error)

 inconsistent: containing discrepancies in codes or names, e.g.,


◼ Age=“42”, Birthday=“03/07/2010”
◼ Was rating “1, 2, 3”, now rating “A, B, C”

 Intentional (e.g., disguised missing data)


◼ Jan. 1 as everyone’s birthday?
How to Handle Missing Data?
8

 Ignore the tuple:


 usually done when class label is missing
 Fill in the missing value manually:
 tedious + infeasible?
 Fill in it automatically with
 A global constant : e.g., “unknown”, a new class?!
 The attribute mean
 The attribute median value
How to Handle Noisy Data?
9

 Binning
 first sort data and partition into (equal-frequency) bins

 e.g., Bin ages of the students of undergraduate

 smooth by bin means, smooth by bin median, etc.

 Regression
 smooth by fitting the data into regression functions

 Clustering
 detect and remove outliers

 Combined computer and human inspection


 detect suspicious values and check by human (e.g., deal with possible outliers)
Binning Methods for Data Smoothing
10

❑ Sorted data for price (in dollars): 4, 8, 9, 15, 21, 21, 24, 25, 26, 28, 29, 34
* Partition into equal-frequency (equi-depth) bins:
- Bin 1: 4, 8, 9, 15
- Bin 2: 21, 21, 24, 25
- Bin 3: 26, 28, 29, 34
* Smoothing by bin means:
- Bin 1: 9, 9, 9, 9
- Bin 2: 23, 23, 23, 23
- Bin 3: 29, 29, 29, 29
* Smoothing by bin boundaries:
- Bin 1: 4, 4, 4, 15
- Bin 2: 21, 21, 25, 25
- Bin 3: 26, 26, 26, 34
Data Integration
11

 Data integration:
 Combines data from multiple sources into a coherent store
 Schema integration: e.g., [Link]-id  [Link]-#
 Integrate metadata from different sources
 Entity identification problem (Name Disambiguation)
 Identify real world entities from multiple data sources, e.g., Bill Clinton = William Clinton
 Detecting and resolving data value conflicts
 Possible reasons:
 different representations: Rs vs. US Dollars
 different scales, e.g., metric vs. British units
Data Reduction Strategies
12

 Data reduction:
 Obtain a reduced representation
 Produces the same (or almost the same)

 Why data reduction?


 Huge volume (terabytes)
 Complex data – difficult to analysis
 Time consuming -

 Data reduction strategies


 Dimensionality reduction, e.g., remove unimportant attributes
 Feature subset selection algorithms
◼ Info Gain
◼ Principal Components Analysis (PCA)
Data Compression
13

Original Data Compressed


Data
lossless

Original Data
Approximated
Data Transformation
14

 A function that maps the entire set of values of a given attribute to a new set of replacement values
s.t. each old value can be identified with one of the new values
 Methods
 Attribute/feature construction
◼ Derived attributes constructed from the given ones
◼ E.g. Age as new attribute instead of Date of Birth
 Normalization:
 Scaled to fall within a smaller, specified range
◼ min-max normalization
Normalization
15

 Min-max normalization: to [new_minA, new_maxA]

v − minA
v' = (new _ maxA − new _ minA) + new _ minA
maxA − minA

 E.g., Let income range $12,000 to $98,000 normalized to [0.0, 1.0]. Then $73,600 is mapped to

73,600 − 12,000
(1.0 − 0) + 0 = 0.716
98,000 − 12,000
16

 PCA
 Info Gain
 Gain Ratio
 Entropy
 Gini
17

You might also like