Classification Algorithms Framework (CAF) to Enable Intelligent Systems Using JetBrains MPS Domain-Specific Languages Environment

Abstract

This paper describes the design and development of a Classification Algorithms Framework (CAF) using the JetBrains MPS domain-specific languages (DSLs) development environment. It is increasingly recognized that the systems of the future will contain some form of adaptivity therefore making them intelligent systems as opposed to the static systems of the past. These intelligent systems can be extremely complex and difficult to maintain. Descriptions at higher-level of abstraction (systemlevel) have long been identified by industry and academia to reduce complexity. This research presents a Framework of Classification Algorithms at system-level that enables quick experimentation with several different algorithms from Naive Bayes to Logistic Regression. It has been developed as a tool to address the requirements of British Telecom's (BT's) data-science team. The tool has been presented at BT and JetBrains MPS and feedback has been collected and evaluated. Beyond the reduction in complexity through the system-level description, the most prominent advantage of this research is its potential applicability to many application contexts. It has been designed to be applicable for intelligent applications in several domains from business analytics, eLearning to eHealth, etc. Its wide applicability will contribute to enabling the larger vision of Artificial Intelligence (AI) adoption in context.

AdaptiveSystems: An Integrated Framework for Adaptive Systems Design and Development Using MPS JetBrains Domain-Specific Modeling Environment

“…At the heart of any adaptive system, data processing through algorithms such as classification algorithms is performed. This is implemented through the CAF DSL language that we developed in our previous publication [10]. The CAF DSL takes as an input a data arff file (this is the type required by the weka libraries implementation that are being called).…”

“…However, it mainly focuses on business aspects without tying them with the corresponding technical implementations. Domainspecific modelling is an emerging area that started from MDE and provides a higher-level of abstraction, isolation of domain-specific aspects with excellent tooling and tailoring to particular domain requirements that include "intelligent" systems (machine learning algorithms, adaptive systems) [8] (DSLs for big datamachine learning) [9].In our previously published work [10] we defined a framework for developing adaptive VLEs. In this paper, we raised the level of abstraction even more by defining an adaptive system language-framework that can be ported to several application domains such as the Adaptive VLE application, the education IoT application, etc.…”

“…The emphasis will be in their reusability and composition. Five main languages were used: AdaptiveSystems (new DSL), Adaptive_VLE (new revised DSL) for collecting data, datamapping (previous publication) [13] for defining the features and their type, datamapping_Adaptive_VLE (new revised DSL) for connecting the datamapping with the Adaptive_VLE application domain language, Classification Algorithms Framework (CAF) (previous publication) [13] language for data processing using a range of classification algorithms. In Section IV, the main benefits from the proposed solution will be detailed with a separate subsection on the applicability to other application domains.…”

AdaptiveVLE: An Integrated Framework for Personalized Online Education Using MPS JetBrains Domain-Specific Modeling Environment

“…This reliefs the educator from having to worry about software development implementation details that comprise the creation of a new VLE. Specifically, in our proposed environment an interface is provided for the educator to configure the data that will be collected, another interface in the same environment to run a classification algorithm from a choice of algorithms using our previously published Classification Algorithms Framework (CAF) DSL [22] and produce the results in the same screen, adapt the learning path for each student according to these results. This led to the creation of a framework for adaptive VLE development (AdaptiveVLE) that will enable educators to focus on their domain issues of collecting and processing data (pieces of learning evidence) leaving implementation details hidden and automatically provided.…”

“…Then data clearing has to be processed manually. Once the set of data to be used for the data processing is identified, the Classification Algorithms Framework (CAF) DSL [5] is utilized and students are categorized according to their achievement levels. According to this information, educators can alter the learning path for individual students to address their individual needs.…”