Mining Big Data for Detecting, Extracting and Recommending Architectural Design Concepts

Abstract

An architecture recommender system can help programmers make better design choices to address their architectural quality attribute concerns while doing their daily programming tasks. We mine big data to detect and extract a large set of architectural design concepts, such as design patterns, design tactics, architecture styles, etc., to be used in our architecture recommender system called ARS. However, mining big data poses many practical challenges for system implementation. The volume, velocity and variety of our data set, like all other big data systems, requires careful planning. This first challenge is to select appropriate technologies from the large number of available products for our system implementation. Building on these technologies our greatest challenge is to custom-fit our algorithms to the parallel processing platform we have selected for ARS, to meet our performance goals.

Designing a Crime-Prevention System by Converging Big Data and IoT

“…Text mining, opinion analysis, social network analysis, and cluster analysis are Big Data analysis techniques [42,43]. We decided that text mining and opinion analysis were the proper techniques for crime-danger analysis, for the reasons listed in Table 6.…”

Tags’ Recommender to Classify Architectural Knowledge Applying Language Models

“…One of the approaches to address the lack of explicit AK at the micro-architecture level is mining issue trackers, version control repositories (generally from open source software projects), or any other document which could contain AK. The main objective is to extract large sets of architectural design concepts [39], or automatically recover design decisions from projects' artifacts [40], which can be textual or based on unified modeling language [41]. However, none of these works consider that in agile environments exist documentation debt; thus, the mining sources used in these works could not exist o could be incomplete or outdated.…”

In Search of Sustainable Design Patterns: Combining Data Mining and Semantic Data Modelling on Disparate Building Data

“…However, using knowledge discovery in data to support future design decisionmaking is an area that is not explored in detail. Studies have explored pattern recognition in simulation data and information extraction from BIM design log files [23], datadriven approaches for energyefficient design by BIM data mining [24], as well as use of data mining for extracting and recommending architectural concepts [25]. Even though these studies demonstrate promising results within the use of KDD for design decision support, they rely on patterns only in design data.…”