With the increase in the usage of databases in various fields and domains, to overcome the challenges in a centralized data mining environment, more and more databases are distributed in networks. The objective of distributed data mining is to perform data mining operations based on the type and availability of distributed resources. To make a proper choice of a particular DDM system/model, the basic differences between each of them must be understood. This paper produces a survey of some of the DDM systems available. It mainly focusses on the homogeneous DDM models. It discusses methods based on semantic web and grid, multi-agent, mobile agent and i-Analyst. A hybrid method AGrIP is also discussed. A comparative analysis is made considering different key issues of DDM. Each method is described in detail by its method/algorithm.
Table of Contents
1. INTRODUCTION
2. Classification of DDM Systems
2.1 Heterogeneous Vs. Homogeneous
2.1.1 Homogeneous DDM systems:
2.1.1.1 DDM systems based on Data Mining Agents
2.1.1.2 DDM models based on Grid
2.1.1.3 Meta-Learning based DDM systems
2.1.2 Heterogeneous Systems
2.1.2.1 DDM models based on CDM
3. Methods & Architecture
3.1 Extendible Multi Agent Data mining System
3.2 CAKE
3.3 i-Analyst based DDM
3.4 Multi Agent DDM model using AATP
3.5 Mobile Agent in DMM
3.6 DDM based on Semantic Web and Grid
3.7 AGrIP based DDM
4. COMPARISON & ANALYSIS
4.1 Comparative Analysis:
4.2 Different approaches
4.3 Challenges
5. CONCLUSION
Objectives and Topics
This paper provides a comprehensive survey of existing Distributed Data Mining (DDM) systems, with a primary focus on homogeneous models, to help researchers and practitioners select the most appropriate DDM architecture based on specific resource and data requirements.
- Architectural classification of DDM systems (Heterogeneous vs. Homogeneous).
- Technical deep-dive into agent-based and grid-based DDM frameworks.
- Comparative analysis of DDM approaches based on performance metrics.
- Evaluation of key challenges in DDM, including data inconsistency and communication costs.
Excerpt from the Book
3.1 Extendible Multi Agent Data mining System
Abbreviated as EMADS, this method is a homogeneous DDM technique. EMADS is a multi-agent driven approach and is advantageous over Agent-driven data mining. The architecture of this model is shown in figure 3. EMADS agents are responsible for accessing local data sources and for collaborative data analysis. EMADS includes data mining agents, data agents, task agents, user agents and mediators.
The data and mining agents are responsible for accessing data and carrying out the data mining process. These agents work in parallel and share information through the task agent. The task agent coordinates the data mining operations, and presents results to the user agent. Mediators are used for agents’ coordination. Data mining is carried out by means of local data mining agents to preserve privacy. Depending on the modes of operation, EMADS can be used by:
• EMADS developers: They develop algorithms
• End Users: Their access is restricted and do data mining tasks
• Contributors: They have restricted access and make the data available for use.
Summary of Chapters
1. INTRODUCTION: Explains the necessity of moving from centralized to distributed databases and defines the core issues like communication cost and knowledge integration.
2. Classification of DDM Systems: Categorizes DDM models into heterogeneous and homogeneous systems, with further sub-classifications based on agents, grid computing, and meta-learning.
3. Methods & Architecture: Provides detailed architectural explanations of specific frameworks like EMADS, CAKE, i-Analyst, AATP, Mobile Agents, Semantic Web/Grid, and AGrIP.
4. COMPARISON & ANALYSIS: Evaluates the different DDM approaches based on criteria such as openness, platform independence, and efficiency, while also highlighting current research challenges.
5. CONCLUSION: Summarizes the findings of the survey and suggests future research directions, such as integrating cloud computing with DDM.
Keywords
DDM, Multi-agent, i-Agent, Ontology, Semantic Web, Grid, CDM, DAP, Distributed Databases, Data Mining, Architecture, Meta-learning, Communication Cost, Knowledge Integration, Distributed Computing
Frequently Asked Questions
What is the primary purpose of this survey?
The paper aims to help developers and researchers understand the differences between various Distributed Data Mining (DDM) systems so they can make informed choices when selecting or designing a model for their specific environment.
What categories of DDM systems are discussed?
The paper primarily classifies systems into Heterogeneous and Homogeneous DDM, with further breakdowns into Agent-based, Grid-based, and Meta-learning-based models.
What is the main objective of the research?
The goal is to provide a detailed survey of current homogeneous DDM models and conduct a comparative analysis to assess their performance based on factors like data integration and fault tolerance.
Which methodology is used to evaluate the models?
The authors use a comparative analysis approach, evaluating each model based on specific performance criteria such as degree of openness, platform independence, communication costs, and final result quality.
What core topics are covered in the main section of the paper?
The main section focuses on the architecture and functionality of specific systems, including EMADS, CAKE, i-Analyst, AATP, mobile agent models, and semantic web-grid hybrid models.
Which keywords best characterize this work?
The work is best characterized by terms like Distributed Data Mining (DDM), multi-agent systems, grid computing, semantic web, and architectural comparative analysis.
How does the i-Analyst framework handle resource management?
i-Analyst uses a two-layer architecture, where the Resource Management Layer handles user interactions and algorithm management, while the Execution Layer (DAP) facilitates agent-based distributed tasks.
What role do PADMAs play in the CAKE architecture?
PADMAs (Parallel Data Mining Agents) enable the system to perform mining tasks in parallel across different data warehouses, significantly reducing execution time.
What are the primary advantages of the EMADS model?
EMADS is a platform-independent, multi-agent driven approach that allows for collaborative data analysis while maintaining local data privacy through the use of local mining agents.
How does AATP address communication efficiency?
The AATP model utilizes a Task Prediction Agent that prioritizes requests and coordinates with the Data Mining Agent to minimize redundant communication across the network.
- Citation du texte
- Swetha Reddy Allam (Auteur), Kotagiri Santhosh (Auteur), 2014, Survey on Distributed Data Mining Systems, Munich, GRIN Verlag, https://www.grin.com/document/294717
