Enviar pesquisa
Carregar
386 390
•
1 gostou
•
243 visualizações
E
Editor IJARCET
Seguir
Tecnologia
Denunciar
Compartilhar
Denunciar
Compartilhar
1 de 5
Baixar agora
Baixar para ler offline
Recomendados
LSTM deep learning method for network intrusion detection system
LSTM deep learning method for network intrusion detection system
IJECEIAES
Survey on Existing Text Mining Frameworks and A Proposed Idealistic Framework...
Survey on Existing Text Mining Frameworks and A Proposed Idealistic Framework...
ijceronline
Automated hierarchical classification of scanned documents using convolutiona...
Automated hierarchical classification of scanned documents using convolutiona...
IJECEIAES
169 s170
169 s170
G H Raisoni College of Engineering Nagpur
Caching on Named Data Network: a Survey and Future Research
Caching on Named Data Network: a Survey and Future Research
IJECEIAES
Text pre-processing of multilingual for sentiment analysis based on social ne...
Text pre-processing of multilingual for sentiment analysis based on social ne...
IJECEIAES
Evidence Data Preprocessing for Forensic and Legal Analytics
Evidence Data Preprocessing for Forensic and Legal Analytics
CSCJournals
Sub1584
Sub1584
International Journal of Science and Research (IJSR)
Recomendados
LSTM deep learning method for network intrusion detection system
LSTM deep learning method for network intrusion detection system
IJECEIAES
Survey on Existing Text Mining Frameworks and A Proposed Idealistic Framework...
Survey on Existing Text Mining Frameworks and A Proposed Idealistic Framework...
ijceronline
Automated hierarchical classification of scanned documents using convolutiona...
Automated hierarchical classification of scanned documents using convolutiona...
IJECEIAES
169 s170
169 s170
G H Raisoni College of Engineering Nagpur
Caching on Named Data Network: a Survey and Future Research
Caching on Named Data Network: a Survey and Future Research
IJECEIAES
Text pre-processing of multilingual for sentiment analysis based on social ne...
Text pre-processing of multilingual for sentiment analysis based on social ne...
IJECEIAES
Evidence Data Preprocessing for Forensic and Legal Analytics
Evidence Data Preprocessing for Forensic and Legal Analytics
CSCJournals
Sub1584
Sub1584
International Journal of Science and Research (IJSR)
Lecture 01 Data Mining
Lecture 01 Data Mining
Pier Luca Lanzi
IRJET- A Study of Privacy Preserving Data Mining and Techniques
IRJET- A Study of Privacy Preserving Data Mining and Techniques
IRJET Journal
Modelling, Conception and Simulation of a Digital Watermarking System based o...
Modelling, Conception and Simulation of a Digital Watermarking System based o...
sipij
COMPLETE END-TO-END LOW COST SOLUTION TO A 3D SCANNING SYSTEM WITH INTEGRATED...
COMPLETE END-TO-END LOW COST SOLUTION TO A 3D SCANNING SYSTEM WITH INTEGRATED...
ijcsit
Ensemble of Probabilistic Learning Networks for IoT Edge Intrusion Detection
Ensemble of Probabilistic Learning Networks for IoT Edge Intrusion Detection
IJCNCJournal
A study of existing ontologies in the io t domain
A study of existing ontologies in the io t domain
Sof Ouni
A Practical Approach To Data Mining Presentation
A Practical Approach To Data Mining Presentation
millerca2
Book of abstract volume 8 no 9 ijcsis december 2010
Book of abstract volume 8 no 9 ijcsis december 2010
Oladokun Sulaiman
Ncct Ieee Software Abstract Collection Volume 1 50+ Abst
Ncct Ieee Software Abstract Collection Volume 1 50+ Abst
ncct
P24120125
P24120125
IJERA Editor
Iaetsd decentralized coordinated cooperative cache
Iaetsd decentralized coordinated cooperative cache
Iaetsd Iaetsd
Iaetsd implementation of chaotic algorithm for secure image
Iaetsd implementation of chaotic algorithm for secure image
Iaetsd Iaetsd
Performance Analysis of Various Data Mining Techniques on Banknote Authentica...
Performance Analysis of Various Data Mining Techniques on Banknote Authentica...
inventionjournals
A new study of dss based on neural network and data mining
A new study of dss based on neural network and data mining
Attaporn Ninsuwan
New Research Articles 2020 June Issue International Journal on Cryptography a...
New Research Articles 2020 June Issue International Journal on Cryptography a...
ijcisjournal
Data Mining With Excel 2007 And SQL Server 2008
Data Mining With Excel 2007 And SQL Server 2008
Mark Tabladillo
Survey of the Euro Currency Fluctuation by Using Data Mining
Survey of the Euro Currency Fluctuation by Using Data Mining
ijcsit
Data Mining
Data Mining
Megha Sharma
A comparative analysis of data mining tools for performance mapping of wlan data
A comparative analysis of data mining tools for performance mapping of wlan data
IAEME Publication
Ijarcet vol-2-issue-4-1579-1582
Ijarcet vol-2-issue-4-1579-1582
Editor IJARCET
Ijarcet vol-2-issue-7-2333-2336
Ijarcet vol-2-issue-7-2333-2336
Editor IJARCET
Ijarcet vol-2-issue-4-1389-1392
Ijarcet vol-2-issue-4-1389-1392
Editor IJARCET
Mais conteúdo relacionado
Mais procurados
Lecture 01 Data Mining
Lecture 01 Data Mining
Pier Luca Lanzi
IRJET- A Study of Privacy Preserving Data Mining and Techniques
IRJET- A Study of Privacy Preserving Data Mining and Techniques
IRJET Journal
Modelling, Conception and Simulation of a Digital Watermarking System based o...
Modelling, Conception and Simulation of a Digital Watermarking System based o...
sipij
COMPLETE END-TO-END LOW COST SOLUTION TO A 3D SCANNING SYSTEM WITH INTEGRATED...
COMPLETE END-TO-END LOW COST SOLUTION TO A 3D SCANNING SYSTEM WITH INTEGRATED...
ijcsit
Ensemble of Probabilistic Learning Networks for IoT Edge Intrusion Detection
Ensemble of Probabilistic Learning Networks for IoT Edge Intrusion Detection
IJCNCJournal
A study of existing ontologies in the io t domain
A study of existing ontologies in the io t domain
Sof Ouni
A Practical Approach To Data Mining Presentation
A Practical Approach To Data Mining Presentation
millerca2
Book of abstract volume 8 no 9 ijcsis december 2010
Book of abstract volume 8 no 9 ijcsis december 2010
Oladokun Sulaiman
Ncct Ieee Software Abstract Collection Volume 1 50+ Abst
Ncct Ieee Software Abstract Collection Volume 1 50+ Abst
ncct
P24120125
P24120125
IJERA Editor
Iaetsd decentralized coordinated cooperative cache
Iaetsd decentralized coordinated cooperative cache
Iaetsd Iaetsd
Iaetsd implementation of chaotic algorithm for secure image
Iaetsd implementation of chaotic algorithm for secure image
Iaetsd Iaetsd
Performance Analysis of Various Data Mining Techniques on Banknote Authentica...
Performance Analysis of Various Data Mining Techniques on Banknote Authentica...
inventionjournals
A new study of dss based on neural network and data mining
A new study of dss based on neural network and data mining
Attaporn Ninsuwan
New Research Articles 2020 June Issue International Journal on Cryptography a...
New Research Articles 2020 June Issue International Journal on Cryptography a...
ijcisjournal
Data Mining With Excel 2007 And SQL Server 2008
Data Mining With Excel 2007 And SQL Server 2008
Mark Tabladillo
Survey of the Euro Currency Fluctuation by Using Data Mining
Survey of the Euro Currency Fluctuation by Using Data Mining
ijcsit
Data Mining
Data Mining
Megha Sharma
A comparative analysis of data mining tools for performance mapping of wlan data
A comparative analysis of data mining tools for performance mapping of wlan data
IAEME Publication
Mais procurados
(19)
Lecture 01 Data Mining
Lecture 01 Data Mining
IRJET- A Study of Privacy Preserving Data Mining and Techniques
IRJET- A Study of Privacy Preserving Data Mining and Techniques
Modelling, Conception and Simulation of a Digital Watermarking System based o...
Modelling, Conception and Simulation of a Digital Watermarking System based o...
COMPLETE END-TO-END LOW COST SOLUTION TO A 3D SCANNING SYSTEM WITH INTEGRATED...
COMPLETE END-TO-END LOW COST SOLUTION TO A 3D SCANNING SYSTEM WITH INTEGRATED...
Ensemble of Probabilistic Learning Networks for IoT Edge Intrusion Detection
Ensemble of Probabilistic Learning Networks for IoT Edge Intrusion Detection
A study of existing ontologies in the io t domain
A study of existing ontologies in the io t domain
A Practical Approach To Data Mining Presentation
A Practical Approach To Data Mining Presentation
Book of abstract volume 8 no 9 ijcsis december 2010
Book of abstract volume 8 no 9 ijcsis december 2010
Ncct Ieee Software Abstract Collection Volume 1 50+ Abst
Ncct Ieee Software Abstract Collection Volume 1 50+ Abst
P24120125
P24120125
Iaetsd decentralized coordinated cooperative cache
Iaetsd decentralized coordinated cooperative cache
Iaetsd implementation of chaotic algorithm for secure image
Iaetsd implementation of chaotic algorithm for secure image
Performance Analysis of Various Data Mining Techniques on Banknote Authentica...
Performance Analysis of Various Data Mining Techniques on Banknote Authentica...
A new study of dss based on neural network and data mining
A new study of dss based on neural network and data mining
New Research Articles 2020 June Issue International Journal on Cryptography a...
New Research Articles 2020 June Issue International Journal on Cryptography a...
Data Mining With Excel 2007 And SQL Server 2008
Data Mining With Excel 2007 And SQL Server 2008
Survey of the Euro Currency Fluctuation by Using Data Mining
Survey of the Euro Currency Fluctuation by Using Data Mining
Data Mining
Data Mining
A comparative analysis of data mining tools for performance mapping of wlan data
A comparative analysis of data mining tools for performance mapping of wlan data
Destaque
Ijarcet vol-2-issue-4-1579-1582
Ijarcet vol-2-issue-4-1579-1582
Editor IJARCET
Ijarcet vol-2-issue-7-2333-2336
Ijarcet vol-2-issue-7-2333-2336
Editor IJARCET
Ijarcet vol-2-issue-4-1389-1392
Ijarcet vol-2-issue-4-1389-1392
Editor IJARCET
Ijarcet vol-2-issue-7-2357-2362
Ijarcet vol-2-issue-7-2357-2362
Editor IJARCET
Ijarcet vol-2-issue-7-2337-2340
Ijarcet vol-2-issue-7-2337-2340
Editor IJARCET
A Unified PID Control Methodology to Meet Plant Objectives
A Unified PID Control Methodology to Meet Plant Objectives
Jim Cahill
Class 19 pi & pd control modes
Class 19 pi & pd control modes
Manipal Institute of Technology
Chapter 8-SAMPLE & SAMPLING TECHNIQUES
Chapter 8-SAMPLE & SAMPLING TECHNIQUES
Ludy Mae Nalzaro,BSM,BSN,MN
Destaque
(8)
Ijarcet vol-2-issue-4-1579-1582
Ijarcet vol-2-issue-4-1579-1582
Ijarcet vol-2-issue-7-2333-2336
Ijarcet vol-2-issue-7-2333-2336
Ijarcet vol-2-issue-4-1389-1392
Ijarcet vol-2-issue-4-1389-1392
Ijarcet vol-2-issue-7-2357-2362
Ijarcet vol-2-issue-7-2357-2362
Ijarcet vol-2-issue-7-2337-2340
Ijarcet vol-2-issue-7-2337-2340
A Unified PID Control Methodology to Meet Plant Objectives
A Unified PID Control Methodology to Meet Plant Objectives
Class 19 pi & pd control modes
Class 19 pi & pd control modes
Chapter 8-SAMPLE & SAMPLING TECHNIQUES
Chapter 8-SAMPLE & SAMPLING TECHNIQUES
Semelhante a 386 390
Mining frequent itemsets (mfi) over
Mining frequent itemsets (mfi) over
IJDKP
Aa31163168
Aa31163168
IJERA Editor
Paper id 25201431
Paper id 25201431
IJRAT
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...
IJERD Editor
Different Classification Technique for Data mining in Insurance Industry usin...
Different Classification Technique for Data mining in Insurance Industry usin...
IOSRjournaljce
A New Data Stream Mining Algorithm for Interestingness-rich Association Rules
A New Data Stream Mining Algorithm for Interestingness-rich Association Rules
Venu Madhav
Mining Maximum Frequent Item Sets Over Data Streams Using Transaction Sliding...
Mining Maximum Frequent Item Sets Over Data Streams Using Transaction Sliding...
ijitcs
FREQUENT ITEMSET MINING IN TRANSACTIONAL DATA STREAMS BASED ON QUALITY CONTRO...
FREQUENT ITEMSET MINING IN TRANSACTIONAL DATA STREAMS BASED ON QUALITY CONTRO...
IJDKP
Fp3111131118
Fp3111131118
IJERA Editor
Mining Stream Data using k-Means clustering Algorithm
Mining Stream Data using k-Means clustering Algorithm
Manishankar Medi
An Improved Differential Evolution Algorithm for Data Stream Clustering
An Improved Differential Evolution Algorithm for Data Stream Clustering
IJECEIAES
1105.1950
1105.1950
Nhat Tam
An Efficient Algorithm for Mining Frequent Itemsets within Large Windows over...
An Efficient Algorithm for Mining Frequent Itemsets within Large Windows over...
Waqas Tariq
Integrating compression technique for data mining
Integrating compression technique for data mining
Dr.Manmohan Singh
Concept Drift Identification using Classifier Ensemble Approach
Concept Drift Identification using Classifier Ensemble Approach
IJECEIAES
A Survey on Data Mining
A Survey on Data Mining
IOSR Journals
Db2425082511
Db2425082511
IJMER
1699 1704
1699 1704
Editor IJARCET
1699 1704
1699 1704
Editor IJARCET
Database techniques for resilient network monitoring and inspection
Database techniques for resilient network monitoring and inspection
TELKOMNIKA JOURNAL
Semelhante a 386 390
(20)
Mining frequent itemsets (mfi) over
Mining frequent itemsets (mfi) over
Aa31163168
Aa31163168
Paper id 25201431
Paper id 25201431
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...
Different Classification Technique for Data mining in Insurance Industry usin...
Different Classification Technique for Data mining in Insurance Industry usin...
A New Data Stream Mining Algorithm for Interestingness-rich Association Rules
A New Data Stream Mining Algorithm for Interestingness-rich Association Rules
Mining Maximum Frequent Item Sets Over Data Streams Using Transaction Sliding...
Mining Maximum Frequent Item Sets Over Data Streams Using Transaction Sliding...
FREQUENT ITEMSET MINING IN TRANSACTIONAL DATA STREAMS BASED ON QUALITY CONTRO...
FREQUENT ITEMSET MINING IN TRANSACTIONAL DATA STREAMS BASED ON QUALITY CONTRO...
Fp3111131118
Fp3111131118
Mining Stream Data using k-Means clustering Algorithm
Mining Stream Data using k-Means clustering Algorithm
An Improved Differential Evolution Algorithm for Data Stream Clustering
An Improved Differential Evolution Algorithm for Data Stream Clustering
1105.1950
1105.1950
An Efficient Algorithm for Mining Frequent Itemsets within Large Windows over...
An Efficient Algorithm for Mining Frequent Itemsets within Large Windows over...
Integrating compression technique for data mining
Integrating compression technique for data mining
Concept Drift Identification using Classifier Ensemble Approach
Concept Drift Identification using Classifier Ensemble Approach
A Survey on Data Mining
A Survey on Data Mining
Db2425082511
Db2425082511
1699 1704
1699 1704
1699 1704
1699 1704
Database techniques for resilient network monitoring and inspection
Database techniques for resilient network monitoring and inspection
Mais de Editor IJARCET
Electrically small antennas: The art of miniaturization
Electrically small antennas: The art of miniaturization
Editor IJARCET
Volume 2-issue-6-2205-2207
Volume 2-issue-6-2205-2207
Editor IJARCET
Volume 2-issue-6-2195-2199
Volume 2-issue-6-2195-2199
Editor IJARCET
Volume 2-issue-6-2200-2204
Volume 2-issue-6-2200-2204
Editor IJARCET
Volume 2-issue-6-2190-2194
Volume 2-issue-6-2190-2194
Editor IJARCET
Volume 2-issue-6-2186-2189
Volume 2-issue-6-2186-2189
Editor IJARCET
Volume 2-issue-6-2177-2185
Volume 2-issue-6-2177-2185
Editor IJARCET
Volume 2-issue-6-2173-2176
Volume 2-issue-6-2173-2176
Editor IJARCET
Volume 2-issue-6-2165-2172
Volume 2-issue-6-2165-2172
Editor IJARCET
Volume 2-issue-6-2159-2164
Volume 2-issue-6-2159-2164
Editor IJARCET
Volume 2-issue-6-2155-2158
Volume 2-issue-6-2155-2158
Editor IJARCET
Volume 2-issue-6-2148-2154
Volume 2-issue-6-2148-2154
Editor IJARCET
Volume 2-issue-6-2143-2147
Volume 2-issue-6-2143-2147
Editor IJARCET
Volume 2-issue-6-2119-2124
Volume 2-issue-6-2119-2124
Editor IJARCET
Volume 2-issue-6-2139-2142
Volume 2-issue-6-2139-2142
Editor IJARCET
Volume 2-issue-6-2130-2138
Volume 2-issue-6-2130-2138
Editor IJARCET
Volume 2-issue-6-2125-2129
Volume 2-issue-6-2125-2129
Editor IJARCET
Volume 2-issue-6-2114-2118
Volume 2-issue-6-2114-2118
Editor IJARCET
Volume 2-issue-6-2108-2113
Volume 2-issue-6-2108-2113
Editor IJARCET
Volume 2-issue-6-2102-2107
Volume 2-issue-6-2102-2107
Editor IJARCET
Mais de Editor IJARCET
(20)
Electrically small antennas: The art of miniaturization
Electrically small antennas: The art of miniaturization
Volume 2-issue-6-2205-2207
Volume 2-issue-6-2205-2207
Volume 2-issue-6-2195-2199
Volume 2-issue-6-2195-2199
Volume 2-issue-6-2200-2204
Volume 2-issue-6-2200-2204
Volume 2-issue-6-2190-2194
Volume 2-issue-6-2190-2194
Volume 2-issue-6-2186-2189
Volume 2-issue-6-2186-2189
Volume 2-issue-6-2177-2185
Volume 2-issue-6-2177-2185
Volume 2-issue-6-2173-2176
Volume 2-issue-6-2173-2176
Volume 2-issue-6-2165-2172
Volume 2-issue-6-2165-2172
Volume 2-issue-6-2159-2164
Volume 2-issue-6-2159-2164
Volume 2-issue-6-2155-2158
Volume 2-issue-6-2155-2158
Volume 2-issue-6-2148-2154
Volume 2-issue-6-2148-2154
Volume 2-issue-6-2143-2147
Volume 2-issue-6-2143-2147
Volume 2-issue-6-2119-2124
Volume 2-issue-6-2119-2124
Volume 2-issue-6-2139-2142
Volume 2-issue-6-2139-2142
Volume 2-issue-6-2130-2138
Volume 2-issue-6-2130-2138
Volume 2-issue-6-2125-2129
Volume 2-issue-6-2125-2129
Volume 2-issue-6-2114-2118
Volume 2-issue-6-2114-2118
Volume 2-issue-6-2108-2113
Volume 2-issue-6-2108-2113
Volume 2-issue-6-2102-2107
Volume 2-issue-6-2102-2107
Último
IAC 2024 - IA Fast Track to Search Focused AI Solutions
IAC 2024 - IA Fast Track to Search Focused AI Solutions
Enterprise Knowledge
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
Martijn de Jong
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
HampshireHUG
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Service
giselly40
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
ThousandEyes
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
ThousandEyes
Evaluating the top large language models.pdf
Evaluating the top large language models.pdf
ChristopherTHyatt
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
Joaquim Jorge
Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivity
Principled Technologies
What Are The Drone Anti-jamming Systems Technology?
What Are The Drone Anti-jamming Systems Technology?
Antenna Manufacturer Coco
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
UK Journal
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024
Rafal Los
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
Safe Software
presentation ICT roal in 21st century education
presentation ICT roal in 21st century education
jfdjdjcjdnsjd
Exploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone Processors
debabhi2
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organization
Radu Cotescu
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
Product Anonymous
How to convert PDF to text with Nanonets
How to convert PDF to text with Nanonets
naman860154
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
Earley Information Science
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc
Último
(20)
IAC 2024 - IA Fast Track to Search Focused AI Solutions
IAC 2024 - IA Fast Track to Search Focused AI Solutions
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Service
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
Evaluating the top large language models.pdf
Evaluating the top large language models.pdf
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivity
What Are The Drone Anti-jamming Systems Technology?
What Are The Drone Anti-jamming Systems Technology?
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
presentation ICT roal in 21st century education
presentation ICT roal in 21st century education
Exploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone Processors
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organization
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
How to convert PDF to text with Nanonets
How to convert PDF to text with Nanonets
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
386 390
1.
ISSN: 2278 –
1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012 ESW-FI: An Improved Analysis of Frequent Itemsets Mining K Jothimani, Dr Antony SelvadossThanamani quickly as time goes by[5]. Therefore, catching the recent Abstract—Frequent itemsets mining play an essential role in trend of data is an important issue when mining frequent many datamining tasks. The frequent itemset mining over data itemsets from data streams. Although the sliding window streams is to find an approximate set of frequent itemsets in model proposed a good solution for this problem, the transaction with respect to a given support and threshold. It should support the flexible trade-off between processing time appearing information of the patterns within the sliding and mining accuracy. It should be time efficient even when the window has to be maintained completely in the traditional user-specified minimum support threshold is small. The approach[11]. objective was to propose an effective algorithm which generates We classify the stream-mining techniques into two frequent patterns in a very less time. Our approach has been categories based on the window model that they adopt in developed based on improvement and analysis of MFI order to provide insights into how and why the techniques are algorithm. In this paper, we introduce a new algorithm ESW-FI, to maintain a dynamically selected set of item sets over useful. First, each element in the datastream can be examined a sliding window. We keep some advantages of the previous only once or twice, making traditional multiple-scan approach and resolve the drawbacks, and produce the approaches infeasible. Second, the consumption of memory improved runtime and memory consumption. The proposed space should be confined in a range, despite that data algorithm gave a guarantee of the output quality and also a elements are continuously streaming into the local site. Third, bound on the memory usage. notwithstanding the data characteristics of incoming stream . may be unpredictable; the mining task should proceed normally and offer acceptable quality of results. Fourth, the Index Terms—Data stream, data-stream mining, Efficient Window, frequent itemset and sliding window. latest analysis result of the data stream should be available as . soon as possible when the user invokes a query. In this paper we consider mining recent frequent itemsets in sliding windowsover data streams and estimate their true I. INTRODUCTION frequencies, while making only one passover the data. In our Data stream is an ordered sequence of elements that arrives design, we actively maintain potentially frequent itemsets ina in timely order. In many application domains, data is compact data structure [8][9]. Compared with existing presented in the form of data streams which originate at some algorithms, our algorithm hastwo contributions as follows: endpoint and are transmitted through the communication 1. It is a real one-pass algorithm. The obsolete channel to the central server. Different from data in transactions are not required whenthey are removed traditional static datasets, data streams are continuous, from the sliding window. unbounded, usually come with high speed and have a data 2. Flexible queries based on continuous transactions in distribution that often changes with time [1][2][3]. It is often the sliding window can be answered with an error refer to as streaming data.. bound guarantee. Some well-known examples include market basket, traffic signals, web-click packets, ATM transactions, and sensor In this paper, we propose a remarkable approximating networks. In these applications, it is desirable that we obtain method for discovering frequent itemsets in a transactional some useful information, like patterns occurred frequently, data stream under the sliding window model. In our design, from the streaming data, to help us make some advanced we actively maintain potentially frequent itemsets ina decision[4]. Data-stream mining is such a technique that can compact data structure. It is a real one-pass algorithm. The find valuable information or knowledge from a great deal of obsolete transactions are not required whenthey are removed primitive data. from the sliding window. Flexible queries based on Recently, the data stream, which is an unbounded continuous transactions in the sliding window can sequence of data elements generated at a rapid rate, provides beanswered with an error bound guarantee. a dynamic environment for collecting data sources. It is likely that the embedded knowledge in a data stream will change II. PRELIMINARIES K. Jothimani is with the Department of Computer Science, NGM A. Related Work College, Pollachi,Tamilnadu,India. Email: jothi1083@yahoo.co.in Frequent-pattern mining has been studied extensively Dr. Antony SelvadossThanamani is with the Department of Computer indata mining, with many algorithms Proposed and Science, NGM College, Pollachi. Tamilnadu, India. Email: selvdoss@yahoo.com implemented. Frequent pattern mining and its . associatedmethods have been popularly used in association All Rights Reserved © 2012 IJARCET 386
2.
ISSN: 2278 –
1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012 rule mining, sequential pattern mining, structured pattern elements are inserted without a deletion. Variable-sized mining, iceberg cube computation, cube gradient analysis, windows have no constraint[18]. associative classification, frequent pattern-basedclustering Fixed-sized and variable-sized windows model several [26], and so on. There are a number of research works which variants of sliding windows. For example, tuple-based windows study the problem of data-stream mining in the first decade of correspond to fixed-sized windows, andtime-based windows 21stcentury. Among these studies, Lossy Counting [3] is the most correspond to variable-sized windows. famous method of mining frequent itemsets (FIs) through data streams under the landmark window model. Besides the In recent two years, a new kind of data-stream mining user-specified minimum support threshold (ms), Lossy Counting method named DSCA has been proposed [12]. DSCA is an also utilizes an error parameter, ε, to maintain those infrequent approximate approach based on the application of the itemsets having the potential to become frequent in the future. Principle of Inclusion and Exclusion in Combinatorial With the use of ε , when an itemset is newly found, Lossy Mathematics [7]. One of the most notable features of DSCA Counting knows the upper bound of counts that itemsets may is that it would approximate the count of an arbitrary itemset, have (in the previous stream data) before it has been monitored through an equation (i.e., Equation (4) in [12]), by only the by the algorithm. sum of counts of the first few orders of its subsets over the B. Previous Algorithms data stream. There are also two techniques named counts bounding and correction, respectively, integrated within Based on the ε mechanism of Lossy Counting, [4] DSCA. The concept of Inclusion and Exclusion Principle [7] is proposed the sliding window method, which can find out valuable that it may also be applied in mining FIs under different frequent itemsets in a data stream under the sliding window window models other than the landmark window. Based on the model with high accuracy. The sliding window method theory of Approximate Inclusion–Exclusion [8], we devise and processes the incoming stream data transaction by propose a new algorithm, called ESW-FI, to discover FIs over transaction.[11][17] Each time when a new transaction is the sliding window in a transactional data stream. inserted into the window, the itemsets contained in that transaction are updated into the data structure incrementally. III. PROBLEM DESCRIPTION Next, the oldest transaction in the original window is dropped Let I = {x1, x2, …,xz} be a set of items (or attributes). An out, and the effect of those itemsets contained in it is also itemset (or a pattern) X is a subset of I and written as X =xi xj…xm. deleted. The sliding window method also has a periodical The length (i.e., number of items) of an itemset X is denoted by operation to prune away unpromising itemsets from its data |X|. A transaction, T, is an itemset and T supports an itemset, X, if structure, and the frequent itemsets are output as mining X⊆T. A transactional data stream is a sequence of continuously result whenever a user requests. incoming transactions. A segment, S, is a sequence of fixed In the sliding window model, there are two typical mining number of transactions, and the size of S is indicated by s. A methods: Moment [5] and CFI-Stream [6]. Both the two methods window, W, in the stream is a set of successive w transactions, aimed at mining Closed Frequent Itemsets (CFIs), a complete where w ≥s. A sliding window in the stream is a window of a and non-redundant representation of the set of FIs. Moment uses fixed number of most recent w transactions which slides forward a data structure called CET to maintain a dynamically selected for every transaction or every segment of transactions. We adopt set of itemsets, which includes CFIs and itemsets that form a the notation Ilto denote all the itemsets of length l together with boundary between CFIs and the rest of itemsets. The CFI-Stream their respective counts in a set of transactions (e.g., over W or S). algorithm, on the other hand, uses a data structure called DIU In addition, we use Tnand Snto denote the latest transaction and tree to maintain nothing other than all Closed Itemsets over the segment in the current window, respectively. Thus, the current sliding window. The current CFIs can be output anytime based window is either W = <Tn-w+1, …,Tn> or W = <Sn-m+1, …, Sn>, on any msspecified by the user. where w and m denote the size of W and the number of segments in W, respectively. Besides, there are still some interesting research works [9] [10] [11] on the sliding window model. In [9] a false-negative In this research, we employ a prefix tree which is organized approach named ESWCA was proposed. By employing a under the lexicographic order as our data structure, and also progressively increasing function of ms, ESWCA greatly processes the growth of itemsets in a lexicographic-ordered reduces the number of potential itemsets and would approximate away. As a result, an itemset is treated a little bit like a sequence the set of FIs over a sliding window. In [10] a data structure (while it is indeed an itemset). A superset of an itemset X is the called DSTreewas proposed to capture information from the one whose length is above |X| and has X as its prefix. We define streams. This tree captures the contents of transactions in a Growthl(X) as the set of supersets of an itemset X whose length window, and arranges tree nodes according to some canonical are l more than that of X, where l ≥0. The number of itemsets in order.According to the overview, one crucial problem is to Growthl(X) is denoted by |Growthl(X)|. efficiently compute the inclusion between two itemsets. This We adopt the symbol cnt(X) to represent the count-value (or costly operation could easily be performed when considering a just count) of an itemset X. The count of X over W, denoted as new representation for items in transactions. From now, each cntW(X), is the number of transactions in W that support X. So item is represented by a unique prime number. cntS(X) represents the count of X over a segment S. Given a A sliding window over a data stream is a bag of last N user-specified minimum support threshold (ms), where 0 <ms≤1, elements ofthe stream. There are two variants of sliding we say that X is a frequent itemset (FI) over W if cntW(X) ms×w, windows based on whether N is fixed(fixed-sized sliding otherwise X is an infrequent itemset (IFI). The FI and IFI over S windows) or variable (variable-sized sliding windows). are defined similarly to those for W. Fixed-sized windows are constrained to perform the insertions Given a data stream in which every incoming transaction has and deletions in pairs, exceptin the beginning when exactly N its items arranged in order, and a changeable value of 387 All Rights Reserved © 2012 IJARCET
3.
ISSN: 2278 –
1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012 msspecified by the user, the problem of mining FIs over a sliding both inserted into and dropped out from the window. The window in the stream is to find out the set of frequent itemsets transaction-by-transaction sliding of a window leads to over the window at different slides. excessively high frequency of processing. In addition, since the We remark that most of the existing stream mining transit of a data stream is usually at a high speed, and the impact methods[13] [14] [15] [19] work with a basic hypothesis that of one single transaction to the entire set of transactions (in the they know the user-specified msin advance, and this current window) is very negligible, making it reasonable to handle the window sliding in a wider magnitude. Therefore, for parameter will remain unchanged all the time before the an incoming transactional data stream to be mined, we propose stream terminates. This hypothesis may be unreasonable, to process on a Segment-oriented window sliding. since in general, a user may wish to tune the value of mseach time he/she makes a query for the purpose of obtaining a more We conceptually divide the sliding window further into preferable mining result. An unchangeable msleads to a serious several, say, m, segments, where the term segment is the one we limitation and may be impractical in most real-life applications. have defined earlier in Section 3. Each of the m segments As a result, we relax this constraint in our problem that the user contains a set of successive transactions and is of the same size s is allowed to change the value of msat different times, while our (i.e., contains the equal number of s transactions). Besides, in method must still work normally. each segment, the summary (which contains I1, I2, and the fair-cutters which we will introduced later) of transactions Using the original count-values of subsets to approximate the belonging to that segment is stored in the data structure we use. count of X may sometimes bring about considerable error. For a We call the sliding of window “segment in-out,” which is better approximation, there is a possible way, which is to defined as follows. boundthe range of counts of subsets for the itemset to be approximated. Definition 1 (Segment in-out) LetScdenote the current segment Let Y be a 3-itemset (to be approximated) and y be a which is going to be inserted into the window next (after it is full 1-subset of Y. To obtain a better approximation of Y, the of s transactions). A segment in-out operation (of the window) is count-value to each subset y with respect to Y has a particular that we first insert Scinto and then extract Sn-m+1 from the original window, where n denotes the id of latest segment in the original range, which can be determined by Y’s 2-subsets that have y window. Therefore, the windows before and after a sliding are W as their common subset, respectively. This range of y’s count is = <Sn-m+1, …,Sn> and W = <Sn-m+2, …, Sn, Sc>, respectively. bounded by an upper bound (i.e., the maximum) and a lower bound (i.e., the minimum), and count-values within this range By taking this segment-based manner of sliding, each time are the set of portions of y’s original count which is more when a segment in-out operation occurs, we delete (or drop out) relevant for y with respect to Y[21]. We define Scby(Y) as the set the earliest segment, which contains the summary of transactions of count-values of y with respect to Y within the range obtained of that segment, from the current window at each sliding. As a through the aforesaid manner. Besides, the upper bound and the result, we need not to maintain the whole transactions within the lower bounds of count-values of y are denoted by uby(Y) and current window in memory all along to support window lby(Y), respectively. sliding[24][25]. In addition, we remark that the parameter m directly affects the consumption of memory. A larger value of m Lemma 1 Let cntub(Y) and cntlb(Y) respectively be the means the window will slide (update) more frequently, while the approximate counts of Y obtained by using uby(Y) and lby(Y)of increasing overhead of memory space is also considerable. In count-values to every 1-subset y ∈Y during the approximating our opinion, an adequate size of m that falls in the range between process. Then cntub(Y) ≤cntlb(Y). 5 and 20 may be suitable for general data streams.[20] Proof: .Let Tub and Tlbbe the sums of counts of 1-subsets of Y Theorem 1 For a 2-itemset X and a threshold ms, let TPub(X) obtained by choosing uby(Y) and lby(Y) for each y ∈ Y, and TPlb(X) be the true-positive rates of X’s 3-supersets in the respectively. Then Tub ≥ Tlbsince uby(Y) ≥ lby(Y) for each y. mining result resulting from adopting Ubc and Lbc to all the According to (1) with the parameters setting m=3 and k=2, we 1-subsets of each superset, respectively. Then we have TPub(X) ≤ can eventually obtain the following simplified equation: cnt(Y)≒ TPlb(X). (1-) c + (d, where the symboljk md denotes the Proof .Let P be the number of FIs of X’s 3-supersets with coefficient of linearly transformed Chebyshev polynomial, and c respect to ms. Also, let Pub and Plbrespectively be the numbers and d represent the sum of counts of Y’s 2-subsets and that of of true FIs of X’s 3-supersets found by choosing Ubc and Lbc counts of Y’s 1-subsets, respectively. Since the value of is to 1-subsets. According to Lemma 1, we have cntub(Y) ≤ less than 1, the coefficient cntlb(Y) for each 3-superset Y of X, which means that the ( for 1-subset term is then negative, which means number of true-positive itemsets (i.e., FIs) obtained by that the value of 1-subset term will be subtracted from the choosing Lbc is at least equal to that of choosing Ubc, i.e., Pub ≤ other term. Thus, by substituting Tuband Tlbfor drespectively Plb. Since TPub(X) = Pub/P and TPlb(X) = Plb/P, we then have in the approximate equation and knowing that Tub ≥Tlb, we TPub(X) ≤TPlb(X). have cntub(Y) ≤ntlb(Y). c Theorem 2 For a 2-itemset X and a threshold ms, let TNub(X) and TNlb(X) be the true-negative rates of X’s 3-supersets in the mining result resulting from adopting Ubc and Lbc to all the IV. EFFICIENT SLIDING-WINDOW PROCESSING 1-subsets of each superset, respectively. Then we have TNub(X) ≥ In research works under the sliding window model [4] [5] [6], TNlb(X). the sliding of window is handled transaction by transaction; however, we have a different opinion. Unlike the landmark Proof .Let N be the number of IFIs of X’s 3-supersets with window model, transactions in the sliding window model will be respect to ms. Also, let Nub and Nlbrespectively be the numbers of All Rights Reserved © 2012 IJARCET 388
4.
ISSN: 2278 –
1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012 true IFIs of X’s 3-supersets determined by choosing Ubc and Lbc phase, which is activated when the window is full of generated to 1-subsets. According to Lemma 1, we have cntub(Y) ≤cntlb(Y) transactions. for each 3-superset Y of X, which means that the number of true-negative itemsets (i.e., IFIs) determined by choosing Ubc is at least equal to that of choosing Lbc, i.e., Nub ≥Nlb. Since TPub(X) = Nub/N and TPlb(X) = Nlb/N, we then have TPub(X) ≥ TPlb(X). Now we discuss the issue of selecting suitable count-values in the bounded range of subsets for approximating an itemset. According to Lemma 1 in Section 3, using the lower bound of counts (Lbc) for 1-subsets always results in a higher approximate count for an itemset than that of using the upper bound of counts (Ubc). It follows from Theorem 1 and Theorem 2 that, adopting Lbc to the 1-subsets to approximate the 3-supersets of an 2-itemset X will reach a higher true-positive rate (i.e., recall ratio) in the result than that of using Ubc, while choosing Ubc to the 1-subsets will obtain a higher true-negative rate (which usually concerns a Fig 5.2: Space Requirement of ESW-FI with T20 higher precision ratio) in the result than that of using Lbc. However, it is actually unknown whether to adopt Ubc or Lbc to1-subsets for approximating an itemset Y. V. EXPERIMENTAL RESULT We performed extensive experiments to evaluate the performance of our algorithmand we present results in this section. We compared our algorithm (ESW-FI)with the MFI algorithm [7].Then we tested the adaptability of ESW-FI by changing the data distribution ofthe dataset. The experiments were performed on a Pentium processor with 4G memory,running Windows 2007 (SP4). Our algorithm is implemented in C# and compiledby using Microsoft Visual Fig 5.3: Execution time of MFI Vs ESW-FI for Updated Studio 2008. In the implementation of three algorithms, weused Transactions the same data structures and subroutines in order to minimize the performancedifferences caused by minor differences. We have usedT20 data set in the experiments. The T20 dataset is generated by the IBM data generator [1]. For T20, the average size of transactions, the average size of the maximal potential frequent itemsets and the numberof items are 20, 4, and 1 000, respectively. Fig 5.1: Execution time of MFI Vs ESW-FI for T20 Fig 5.4: Space Requirement between MIF Vs ESW-FI with more number of transactions Comparison between MFI and ESW-FI basedon different block size to observe the influence of different block sizes on algorithm, we take 2×104, 5×104and 10 × 104 as an example. Meanwhile, the size of the window is fixed 5 × 105.In this case, this window contains 25, 10 and 5 blocks, respectively. The data sets were broken into batches of10K size transactions and provided to our program through standard The experimental results of the execution time and the input.For ESW-FI, the whole procedure can be divided into two space requirement are plotted in Figure 5.1 to 5.4, different phases namely, window initialization phase, which is respectively. We collected the total number of seconds and activated when the number of transactions generated so far is not the size of storage space required in KB per 10 × 5 more than the size of the sliding window and windowsliding 389 All Rights Reserved © 2012 IJARCET
5.
ISSN: 2278 –
1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012 transactions. These results basically keep stable as the [18] Y. Chi, H. Wang, P.S. Yu, & R.R. Muntz, “Moment: maintaining closed frequent itemsets over a stream sliding window”, Proc. 4th window moves forward. The above experimental results IEEE Conf. on Data Mining, Brighton, UK, 2004, pp. 59–66. provide evidence that two algorithm can handle long data [19] N. Jiang & L. Gruenwald, “CFI-Stream: mining closed frequent streams both. As ESW-FI may keep multiple triples for one Itemsets in data streams”, Proc. 12th ACM SIGKDD Conf. on Knowledge Discovery and Data Mining, Philadelphia, PA,USA, 2006, significant itemset, it needs more memory and time than MFI pp. 592–597. does. [20] Quest Data Mining Synthetic Data Generation Code. Available: http://www.almaden.ibm.com/cs/projects/iis/hdb/Projects/data_minin VI. CONCLUSION g/datasets/syndata.html [21] P. Indyk, D. Woodruff, “Optimal approximations of the frequency We mainly discuss how to discover recent frequent moments of data streams”, Proceedings of the thirty-seventh annual itemsets in sliding windows over data streams. An efficient ACM symposium on Theory of computing, pp.202–208, 2005. [22] H.F Li, S.Y. Lee, M.K. Shan, “An Efficient Algorithm for Mining algorithm is presented in detail. Compared with previous Frequent Itemsets over the Entire History of Data Streams”, In algorithms, this algorithm doesn’t keep the data in the Proceedings of First International Workshop on Knowledge Discovery window, which considerably increase the scalability of the in Data Streams 9IWKDDS, 2004. [23] H.F Li, S.Y. Lee, M.K. Shan, “Online Mining (Recently) Maximal algorithm. Moreover, it can figure out answers with an error Frequent Itemsets over Data Streams”, In Proceedings of the 15th bound guarantee for continuous transactions in the window. IEEE International Workshop on Research Issues on Data Engineering The extensive experiment results demonstrate the (RIDE), 2005. [24] Lin C.-H., Chiu D.-Y., Wu Y.-H. and Chen A.L.P.: Mining Frequent effectiveness and efficiency of our approach. We can Itemsets from Data Streams with a Time-Sensitive Sliding Window. In implement this algorithm in traffic networks especially for Proc SIAM International Conference on Data Mining. (2005). mining the itemsets in high speed streams. [25] Ho, C. C., Li, H. F., Kuo, F. F., & Lee, S. Y. (2006). Incremental mining of sequential patterns over a stream sliding window. In Proceedings of IEEE international workshop on mining evolving and REFERENCES streaming data. [26] K Jothimani, Dr Antony SelvadossThanmani, “MS: Multiple Segments [1] R. Agrawal, T. Imielinski, and A. Swami. Mining Association Rules with Combinatorial Approach for Mining Frequent Itemsets Over Data between Sets of Items in Large Databases. In Proceedings of the 1993 Streams”, IJCES International Journal of Computer Engineering International Conference on Management of Data, pp. 207-216, 1993. Science, Volume 2 Issue 2 ISSN : 2250:3439 [2] R. Agrawal and R. Srikant. Fast Algorithms for Mining Association [27] K Jothimani, Dr Antony SelvadossThanmani, “An Algorithm for Rules. In Proceedings of the 20th International Conference on Very Mining Frequent Itemsets,” “International Journal for Computer Large Data Bases, pp. 487-499, 1994 Science Engineering and Technology IJCSET, March 2012,Vol 2, [3] J.H. Chang & W.S. Lee, “A sliding window method for finding recently Issue 3,1012-1015. frequent itemsets over online data streams’, Journal of Information Science and Engineering, 20(4), 2004, pp. 753–762 [4] Y. Zhu & D. Shasha, “Stat Stream: statistical monitoring of thousands of data streams in real time”, Proc. 28th Conf. on Very Large Data K Jothimanireceived her Bachelor degree in Computer Science from Bases, Hong Kong, China, 2002, pp. 358–369. Bharathidasan University, Trichy in 2003. She received her Master degree in [5] G.S. Manku& R. Motwani, “Approximate frequency counts over data Computer Applications in 2008. She pursued her Master of Philosophy in streams”, Proc. 28th Conf. on Very Large Data Bases, Hong Kong, Computer Science in the year 2009 from Vinayaka Missions University, China, 2002, pp. 346–357. Salem. Currently she is a research scholar of the Department of Computer [6] J.H. Chang & W.S. Lee, “A sliding window method for finding recently Science, NGM College under Bharathiyar University, Coimbatore. She had frequent itemsets over online data streams”, Journal of Information five years of experience in the computer field in technical as well as science and Engineering, 20(4), 2004, pp. 753–762. non-technical. She is a life member of Indian Society for Technical [7] M.N. Garofalakis, J. Gehrke, & R. Rastogi, Querying and mining data Education from the year 2009. Also she is active member of Computer streams: you only get one look (A Tutorial), Proc. 2002 ACM Society of India (CSI). She has published more than ten papers in SIGMOD Conf. on Management of Data, Madison, Wisconsin, 2002, international and national conferences including international journals. Her p. 635. area of interests includes Data mining, Knowledge Engineering and Image [8] Y. Zhu & D. Shasha, “Stat Stream: statistical monitoring of thousands Processing. of data streams in real time”, Proc. 28th Conf. on Very Large Data Email: jothi1083@yahoo.co.in Bases, Hong Kong, China, 2002, pp. 358–369. . [9] G.S. Manku& R. Motwani, “Approximate frequency counts over data Dr Antony SelvadossThanamaniis presently working as professor and streams”, Proc. 28th Conf. on Very Large Data Bases, Hong Kong, Head, Dept of Computer Science, NGM College, Coimbatore, China, 2002, pp. 346–357. India(affiliated to Bharathiar University, Coimbatore). He has published [10] J.H. Chang & W.S. Lee, “A sliding window method for finding recently many papers in international/national journals and written many books. His frequent itemsets over online data streams”, Journal of Information areas of interest include E-Learning, Software Engineering, Data Mining, science and Engineering, 20(4), 2004, pp. 753–762. Networking, Parallel and Distributed Computing. He has to his credit 25 [11] J. Cheng, Y. Ke, & W. Ng,” Maintaining frequent itemsets over years of teaching and research experience. His current research interests high-speed data streams”, Proc. 10th Pacific-Asia Conf. on include Grid Computing, Cloud Computing, Semantic Web. He is a life Knowledge Discovery and Data Mining, Singapore, 2006, pp.462–467. member of Computer Society of India, Life member of Indian Society for [12] C.K.-S. Leung & Q.I. Khan, “DSTree: a tree structure for the mining Technical Education, Life member of Indian Science Congress, Life member of frequent sets from data streams,” Proc. 6th IEEE Conf. on Data of Computer Science, Teachers Associates, Newyork. Mining, Hong Kong, China, 2006, pp. 928–932. Email:-selvdoss@yahoo.com [13] B. Mozafari, H. Thakkar, & C. Zaniolo, “Verifying and mining frequent patterns from large windows over data streams”, Proc. 24th Conf. on Data Engineering, Mexico, 2008, pp. 179–188. [14] K.-F. Jea& C.-W. Li, “Discovering frequent itemsets over transactional data streams through an efficient and stable approximate approach, Expert Systems with Applications”, 36(10), 2009, pp. 12323–12331. [15] F. Bodon, “A fast APRIORI implementation”, Proc. ICDM Workshop on Frequent Itemset Mining Implementations (FIMI’03), 2003. [16] N. Jiang and L. Gruenwald, “Research Issues in Data Stream Association Rule Mining”. In SIGMOD Record, Vol. 35, No. 1, Mar. 2006. [17] Frequent Itemset Mining Implementations Repository (FIMI).Available: http://fimi.cs.helsinki.fi/ All Rights Reserved © 2012 IJARCET 390
Baixar agora