M. Alnoukari 1 Z.Alzoabi 2 S.Hanna 1 1 Arab International University , Damascus, Syria.

1. Applying Adaptive Software Development (ASD) Agile Modeling on Predictive Data Mining Applications: ASD-DM Methodology M. Alnoukari1Z.Alzoabi2S.Hanna1 1 Arab International University , Damascus, Syria. 2 Arab Academy for Banking and Financial Sciences, Damascus, Syria

2. Plan • Introduction • Problem • What Characterizes Data Mining Applications? • CRISP-DM methodology. • Applying ASD method on predictive data mining applications: ASD-DM Methodology • A Data Mining Case Study in Automotive Manufacturing Domain • Conclusion and Future Works

3. Introduction • The idea of applying data mining techniques on software engineering data has existed since mid 1990s. • Data mining techniques are applied to: • Analyze the problems raised during the life cycle of a software project development. • Determine if two software components are related or not. • Software maintenance • Software testing • Software reliability analysis, and software quality. • Many questions arise when trying to apply data mining techniques on software engineering field: • What types of SE data are available to be mined? • Which SE tasks can be held using data mining? • How are data mining techniques used in SE?

4. Problem • The world becomes increasingly dynamic, the traditional static modeling may not be able to deal with it. • Data mining applications require greater diversity of technology, business skills, and knowledge than the typical applications. • One solution is to use agile modeling that is characterized with flexibility and adaptability. • We propose a framework named ASD-DM based on Adaptive Software Development (ASD) that can easily adapt with predictive data mining applications. • A case study in automotive manufacturing domain was explained and experimented to evaluate ASD-DM methodology.

5. WhatCharacterizes DM Applications? • Data mining applications are characterized by the ability to deal with the explosion of business data and accelerated market changes. • These characteristics help providing powerful tools for decision makers. • Such tools can be used by business users (not only PhDs, or statisticians) for analyzing huge amount of data for patterns and trends . • The most widely used methodology when applying data mining processes is named CRISP-DM.

6. CRISP-DM

7. ASD-DM Methodology • ASD (Adaptive Software Development) modeling replaces the static Plan-Design-Build lifecycle, with the dynamic Speculate-Collaborate-Learn life cycle. • “Speculation” recognizes the uncertain nature of complex problems such as predictive data mining. • “Collaboration” among different stakeholders, in order to improve their decision making ability. • “Learning” component in order to test knowledge raised by practices iteratively after each cycle.

8. ASD-DM Methodology

9. Case Study: Automotive Manufacturing Domain • The information gathered in order to produce automotive data mining solution are the following: • Supply chain process (sales, inventory, orders, production plan). • Manufacturing information (car configurations/packages/options codes and description). • The main goal is get some initial positive results on prediction and to measure the prediction score of different data sources using findings of correlation studies.

10. Prediction Results Using Neural Networks

11. Prediction Results Using Linear Regression

12. Conclusion & Future Work • ASD-DM framework ensures continuous learning, and intense collaboration among developers, testers, and customers. • Future work: • How can ASD-DM framework enhance knowledge sharing and organizational learning. • How can ASD-DM framework help organizations achieving their business strategy.