Tailoring the Engineering Design Process Through Data and Process Mining

ABSTARCT :

Engineering changes (ECs) are new product devel-opment activities addressing external or internal challenges, suchas market demand, governmental regulations, and competitivereasons. The corresponding EC processes, although perceived asstandard, can be very complex and inefficient. There seem to besignificant differences between what is the “officially” documentedand the executed process. To better understand this complexity, wepropose a data-driven approach, based on advanced text analyticsand process and data mining techniques. Our approach sets the firststeps toward an automatic analysis, extracting detailed events froman unstructured event log, which is necessary for an in-depth un-derstanding of the EC process. The results show that the predictiveaccuracy associated with certain EC types is high, which assures themethod applicability. The contribution of this article is threefold:1) a detailed model representation of the actual EC process isdeveloped, revealing problematic process steps (such as bottleneckdepartments); 2) homogeneous, complexity-based EC types aredetermined (ranging from “standard” to “complex” processes);and 3) process characteristics serving as predictors for EC typesare identified (e.g., the sequence of initial process steps determinesa “complex” process). The proposed approach facilitates process and product innovation, and efficient design process management in future projects.

EXISTING SYSTEM :

? The main idea of business process mining is to extract the execution of business processes recorded in event logs available in today’s information system in order to automatically construct the models of business processes, compare existing business process models with the new automatically constructed models to identify deviations and bottlenecks and to enhance the business processes. ? Over the last decade, the amount of event data has greatly increased, and process mining techniques have significantly matured. ? An automatic process modelling methodology that takes an event log where the execution of existing processes is recorded as input and produces a business process model as output.

DISADVANTAGE :

? One explanation for this lack of research is the challenges involvedin analyzing the inherently sequential structure of process data. ? While one event can have a short cycle time in one process stage,the same event may have longer cycle times in a subsequentstage ? Although this information can be useful to determine process performance,it lacks the description of the actual tasks, and the types of employees and departments involved in the change project. ? More importantly, this information provides only limited insights intothe interactions that take place.

PROPOSED SYSTEM :

• In education, data mining techniques, such as classification and clustering, are often used to categorize (or profile) students based on the kinds of personal learning data described in the section on the research base, on student demographic data, or both. • Kardan and Conati (2011) proposed a user modeling framework that relies on interaction logs to identify different types of learners, as well as their characteristic interactions with the learning system. • Any predictive models proposed for consequential use (such as assigning students to services or qualifying them for advanced courses) should be transparent and backed up by solid empirical evidence based on data from similar institutions.

ADVANTAGE :

• It is important to not only examine the number of rulesand coverage ratio of positive and negative cases, but also theperformance of the JRip rule-based classification model. • This can be assessed by evaluating the performance of theclassification model on the remaining 30% of data (1498-testing dataset), and by carrying out a tenfold cross validation. • K-fold cross validation is a model validation technique for as-sessing how the results of applying a classifier will generalize foran independent dataset • The usual performance measuresare considered, namely, accuracy, error, precision, recall, andF-measure.