Armen Zakarian

This paper presents a new associative classification algorithm for data mining. The algorithm uses elementary set concepts, information entropy and database manipulation techniques to develop useful relationships between input and output attributes of large databases. These relationships (knowledge) are represented using IF–THEN association rules, where the IF portion of the rule includes a set of input attributes features and THEN portion of the rule includes a set of output attributes that represent decision outcome. Application of the algorithm is presented with a thermal spray process control case study. Thermal spray is a process of forming a desired shape of material by spraying melted metal on a ceramic mould. The goal of the study is to identify spray process input parameters that can be used to effectively control the process with the purpose of obtaining better characteristics for the sprayed material. Detailed discussion on the source and characteristics of the data sets is also presented.

Modelling and analysis of system robustness is of interest to researchers and practitioners for several reasons. Robust systems can minimize product development time, warranty costs, and improve product quality. The problem of attaining system robustness is often not addressed formally because design teams work independently of each other when designing system components. In this paper, a framework is presented for robust system development based upon system modelling, integration analysis, and quality engineering techniques. System robustness is achieved by specifying sub-system configurations within the overall system to minimize subsystem-to-subsystem interactions and overall system sensitivity to noise factors. System modelling represents system structure, captures system requirements and identifies required functional interfaces among system components. Once the interfaces among system components are identified, a component—component interaction matrix of interfaces is developed. Integration analysis then groups system components into subsystems and identifies interfaces (main elements of robustness study) between subsystems. Experimental design techniques are used to identify the combinations of levels of these interfaces, if any, that result in a robust system specification. This identified combination of levels then is incorporated into designing product subsystems and components. The application of the approach is illustrated with an automotive example of the design of a vehicle seat system. Discussion on the importance of system modelling and traditional robust design procedures and on the difficulties associated with the development of robust systems is also presented.

A new association rule-generation algorithm is presented for mining automotive warranty data. The algorithm uses elementary set concept and database manipulation techniques to develop useful relationships between product attributes and causes of failure. These relationships (knowledge) are represented using IF–THEN association rules, where the IF portion of the rule includes set of attributes representing product features (e.g. production date, repair date, mileage-at-repair, transmission, engine type, etc.) and the THEN portion of the rule includes set of attributes that represent decision outcome (e.g. problem-related labor code). Once association rules are developed, the algorithm applies a statistical analysis technique to evaluate the significance of each rule. The rules that pass the significance test are reported in a solution. Application of the association rule-generation algorithm is presented with a data-mining case study from the automotive industry. The knowledge (rules) extracted from the automotive warranty data are used to identify root causes of a particular warranty problem or to develop useful conclusions. Detailed discussion on the source and characteristics of warranty data is also presented.

One of the key issues to business process control is the identification of measurable process attributes. For manufacturing processes these are typically physical parameters of the process (e.g. temperature, set points) or physical attributes of the manufactured product (e.g. dimension, functional performance). However, for business processes the metrics are more The challenge has been to develop metrics that capture the contributing subtle and hard to measure factors for business process control. This paper presents an analytical model that uses the weights-of-evidence concept to convert answers to audit or self-assessment questions into a single numerical process quality index. This index is used to forecast process success or failure and monitor its performance from start to end. The application of the approach is illustrated with an automotive industry product development sub-process where the process performance metric is the field warranty data, i.e. incidents per thousand vehicles (IPTV). The analytical model converts process self-assessment (failure mode and effect analysis) questions into a single numeric process quality index. The validity of the model is reflected in the strength of the correlation between the index and the IPTV results. Also, in this paper a measure is developed for identifying critical process quality assessment questions. This measure quantifies the deviation in the automotive business process that should have more focus. The significance of the analytical model proposed in this research is that the project managers or quality assurance auditors may be able to use the metric to predict product quality at any point in the product development process.