Agent based preprocessing

Research output: Chapter in Book/Report/Conference proceedingConference contribution

9 Citations (Scopus)

Abstract

The current data mining tools is used to build knowledge based on a huge historical data. At present, businesses are facing with fast growing data that are very valuable in contributing knowledge. Knowledge should be updated regularly in order to ensure its quality and precision thus improve the decision making process. Data mining has shown great potential in extracting valuable knowledge from large databases. However, current data mining algorithms and tools are costly and several are too complex in their operations when dealing with large databases. In recent years, agents have become a popular paradigm in computing, because its autonomous, flexible and provides intelligence. Embedding agents in the current data mining processes and tools are believed to be able to solve the obstacle. One of the most important process in data mining is data preprocessing. It is reported that 60% of the data mining project is on preprocessing. Data preprocessing involves integration, selection, cleaning and transformation of data set that will be used for mining. This paper focuses on an agent-based preprocessing framework. The aims is to provides an auto preprocessing a set of new data, which suite to data mining novice user. The proposed agent based preprocessing framework consists of seven agents: User Interface agents, Coordinator Agent, Identify Agent, CleanMiss Agent, CleanNoisy Agent, Transformation Agent and Discretization Agent. User Interface Agent is designed in such a way to provide interface suite to novice users. Coordinator agent is responsible for coordinating and cooperating with all other agents to achieve the goals. Identify agent responsible to provide an adaptive user data cleaning profiling. CleanMiss Agent, CleanNoisy Agent, Transformation Agent and Discretization Agent provide various types of techniques autonomously, which ended with proposing the best cleaning techniques from various types of techniques to keep in the preprocessing profile. This paper is start by introducing the data mining process problem includes data preprocessing which agent can solve data mining problems. By applying agent in data preprocessing, a tool that intelligence yet flexible can be produced.

Original languageEnglish
Title of host publication2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007
Pages219-223
Number of pages5
DOIs
Publication statusPublished - 2007
Event2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007 - Kuala Lumpur
Duration: 25 Nov 200728 Nov 2007

Other

Other2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007
CityKuala Lumpur
Period25/11/0728/11/07

Fingerprint

Data mining
Cleaning
User interfaces
Decision making
Industry

Keywords

  • Agents
  • Data mining
  • Data preprocessing

ASJC Scopus subject areas

  • Artificial Intelligence
  • Control and Systems Engineering

Cite this

Ali Othman, Z., Abu Bakar, A., Hamdan, A. R., Omar, K., & Mohd Shuib, N. L. (2007). Agent based preprocessing. In 2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007 (pp. 219-223). [4658378] https://doi.org/10.1109/ICIAS.2007.4658378

Agent based preprocessing. / Ali Othman, Zulaiha; Abu Bakar, Azuraliza; Hamdan, Abdul Razak; Omar, Khairuddin; Mohd Shuib, Nor Liyana.

2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007. 2007. p. 219-223 4658378.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ali Othman, Z, Abu Bakar, A, Hamdan, AR, Omar, K & Mohd Shuib, NL 2007, Agent based preprocessing. in 2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007., 4658378, pp. 219-223, 2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007, Kuala Lumpur, 25/11/07. https://doi.org/10.1109/ICIAS.2007.4658378
Ali Othman Z, Abu Bakar A, Hamdan AR, Omar K, Mohd Shuib NL. Agent based preprocessing. In 2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007. 2007. p. 219-223. 4658378 https://doi.org/10.1109/ICIAS.2007.4658378
Ali Othman, Zulaiha ; Abu Bakar, Azuraliza ; Hamdan, Abdul Razak ; Omar, Khairuddin ; Mohd Shuib, Nor Liyana. / Agent based preprocessing. 2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007. 2007. pp. 219-223
@inproceedings{783eef6357494d1ea4765af5f66b5425,
title = "Agent based preprocessing",
abstract = "The current data mining tools is used to build knowledge based on a huge historical data. At present, businesses are facing with fast growing data that are very valuable in contributing knowledge. Knowledge should be updated regularly in order to ensure its quality and precision thus improve the decision making process. Data mining has shown great potential in extracting valuable knowledge from large databases. However, current data mining algorithms and tools are costly and several are too complex in their operations when dealing with large databases. In recent years, agents have become a popular paradigm in computing, because its autonomous, flexible and provides intelligence. Embedding agents in the current data mining processes and tools are believed to be able to solve the obstacle. One of the most important process in data mining is data preprocessing. It is reported that 60{\%} of the data mining project is on preprocessing. Data preprocessing involves integration, selection, cleaning and transformation of data set that will be used for mining. This paper focuses on an agent-based preprocessing framework. The aims is to provides an auto preprocessing a set of new data, which suite to data mining novice user. The proposed agent based preprocessing framework consists of seven agents: User Interface agents, Coordinator Agent, Identify Agent, CleanMiss Agent, CleanNoisy Agent, Transformation Agent and Discretization Agent. User Interface Agent is designed in such a way to provide interface suite to novice users. Coordinator agent is responsible for coordinating and cooperating with all other agents to achieve the goals. Identify agent responsible to provide an adaptive user data cleaning profiling. CleanMiss Agent, CleanNoisy Agent, Transformation Agent and Discretization Agent provide various types of techniques autonomously, which ended with proposing the best cleaning techniques from various types of techniques to keep in the preprocessing profile. This paper is start by introducing the data mining process problem includes data preprocessing which agent can solve data mining problems. By applying agent in data preprocessing, a tool that intelligence yet flexible can be produced.",
keywords = "Agents, Data mining, Data preprocessing",
author = "{Ali Othman}, Zulaiha and {Abu Bakar}, Azuraliza and Hamdan, {Abdul Razak} and Khairuddin Omar and {Mohd Shuib}, {Nor Liyana}",
year = "2007",
doi = "10.1109/ICIAS.2007.4658378",
language = "English",
isbn = "1424413559",
pages = "219--223",
booktitle = "2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007",

}

TY - GEN

T1 - Agent based preprocessing

AU - Ali Othman, Zulaiha

AU - Abu Bakar, Azuraliza

AU - Hamdan, Abdul Razak

AU - Omar, Khairuddin

AU - Mohd Shuib, Nor Liyana

PY - 2007

Y1 - 2007

N2 - The current data mining tools is used to build knowledge based on a huge historical data. At present, businesses are facing with fast growing data that are very valuable in contributing knowledge. Knowledge should be updated regularly in order to ensure its quality and precision thus improve the decision making process. Data mining has shown great potential in extracting valuable knowledge from large databases. However, current data mining algorithms and tools are costly and several are too complex in their operations when dealing with large databases. In recent years, agents have become a popular paradigm in computing, because its autonomous, flexible and provides intelligence. Embedding agents in the current data mining processes and tools are believed to be able to solve the obstacle. One of the most important process in data mining is data preprocessing. It is reported that 60% of the data mining project is on preprocessing. Data preprocessing involves integration, selection, cleaning and transformation of data set that will be used for mining. This paper focuses on an agent-based preprocessing framework. The aims is to provides an auto preprocessing a set of new data, which suite to data mining novice user. The proposed agent based preprocessing framework consists of seven agents: User Interface agents, Coordinator Agent, Identify Agent, CleanMiss Agent, CleanNoisy Agent, Transformation Agent and Discretization Agent. User Interface Agent is designed in such a way to provide interface suite to novice users. Coordinator agent is responsible for coordinating and cooperating with all other agents to achieve the goals. Identify agent responsible to provide an adaptive user data cleaning profiling. CleanMiss Agent, CleanNoisy Agent, Transformation Agent and Discretization Agent provide various types of techniques autonomously, which ended with proposing the best cleaning techniques from various types of techniques to keep in the preprocessing profile. This paper is start by introducing the data mining process problem includes data preprocessing which agent can solve data mining problems. By applying agent in data preprocessing, a tool that intelligence yet flexible can be produced.

AB - The current data mining tools is used to build knowledge based on a huge historical data. At present, businesses are facing with fast growing data that are very valuable in contributing knowledge. Knowledge should be updated regularly in order to ensure its quality and precision thus improve the decision making process. Data mining has shown great potential in extracting valuable knowledge from large databases. However, current data mining algorithms and tools are costly and several are too complex in their operations when dealing with large databases. In recent years, agents have become a popular paradigm in computing, because its autonomous, flexible and provides intelligence. Embedding agents in the current data mining processes and tools are believed to be able to solve the obstacle. One of the most important process in data mining is data preprocessing. It is reported that 60% of the data mining project is on preprocessing. Data preprocessing involves integration, selection, cleaning and transformation of data set that will be used for mining. This paper focuses on an agent-based preprocessing framework. The aims is to provides an auto preprocessing a set of new data, which suite to data mining novice user. The proposed agent based preprocessing framework consists of seven agents: User Interface agents, Coordinator Agent, Identify Agent, CleanMiss Agent, CleanNoisy Agent, Transformation Agent and Discretization Agent. User Interface Agent is designed in such a way to provide interface suite to novice users. Coordinator agent is responsible for coordinating and cooperating with all other agents to achieve the goals. Identify agent responsible to provide an adaptive user data cleaning profiling. CleanMiss Agent, CleanNoisy Agent, Transformation Agent and Discretization Agent provide various types of techniques autonomously, which ended with proposing the best cleaning techniques from various types of techniques to keep in the preprocessing profile. This paper is start by introducing the data mining process problem includes data preprocessing which agent can solve data mining problems. By applying agent in data preprocessing, a tool that intelligence yet flexible can be produced.

KW - Agents

KW - Data mining

KW - Data preprocessing

UR - http://www.scopus.com/inward/record.url?scp=57949107047&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=57949107047&partnerID=8YFLogxK

U2 - 10.1109/ICIAS.2007.4658378

DO - 10.1109/ICIAS.2007.4658378

M3 - Conference contribution

SN - 1424413559

SN - 9781424413553

SP - 219

EP - 223

BT - 2007 International Conference on Intelligent and Advanced Systems, ICIAS 2007

ER -