Accumulative frame models for extraction and aggregation of data for knowledge and processes in education

Статията представя резултатите от изследване за приложение на фрейми за представяне на знания и процеси в е-обучението. Класическата нотация на фреймови модели е доразвита в т.нар акумулативен фреймов модел. Представени са създадени на негова база информационни и компютърни модели и софтуерни средства за извличан и агрегиране на данни. Обсъдени са и резултатите от проведеното апробиране и експериментиране за автоматизация на задачи за електронно обучение.

The paper presents a part of a study on the application of frames for the representation of knowledge and processes in programming training. The proposed accumulative frame model serves as a basis for automation of e-learning and e-training activities by the implementation of tools for extraction, aggregation and accumulation of data and knowledge for educational needs. The created on its basis information and computer models and a software tool for data extraction and aggregation are depicted.

The paper presents a part of a study conducted for the needs of e-learning on the application of frames for representation of knowledge and processes in training on a particular subject domain. As a result of the study the classical notion of the frame models is developed further and the so-called accumulative frame model is proposed. The main rationales for the introduction of the accumulative frame model are: (1) to provide a formalism for conceptual modelling in subject domains and (2) to serve as a basis for automation of e-learning and e-training activities by implementation of intelligent tools for extraction, aggregation and accumulation of data and knowledge for educational needs. A system of 36 accumulative frame models and its verification for fulfilment of the intended purpose in the rationale (1) are presented.

A software environment CumuFormo for designing prototypes of frames and instances of frames with slots of different types is introduced. CumuFormo supports different databases and facilitates modeling of pedagogical scenarios in the form of e-learning workflows. The software tool Alpaca and a workflow interpreter, named EMS, are used in the development of CumuFormo.

Използването на фризови представяния на знания в обучението има дълга история. Статията представя част от изследване за приложение на фреймови представяния на знания за нуждите на е-обучението. На базата на предложения акумулативен фреймов модел за представяне и акумулиране на знания е създадена система от 20 акумулативни фреймови модела за концептуално представяне на знания в областта на програмирането. Предложени за 4 основни типа задачи за изучаване на синтаксиса на оператори на езика за програмиране С++.

The use of frame-based knowledge representations in training methods has a long history. The paper presents a part of a study on the application of frame-based computer representations of knowledge for the needs of e-learning. It proposes an innovative modification of frame models, called accumulative frame model (AFM). The AFM is a development of the classical understanding of a frame model, and therefore has its all typical features, but beside the possibilities for knowledge representation, it adds to the model the means for knowledge accumulation also. This provides а particular flexibility in the implementation of processes for automated extraction and aggregation of data and knowledge for training needs. In addition, a model of a system of AFM is introduced, with the purpose of conceptual representation of knowledge in separate subject domains. The proposed models are approbated for conceptual modelling of a particular subject domain of study-the domain of computer programming.

The generality of the frame representations determines their broad application in a lot of diverse areas. The article introduces a part of a research on the implementation of the frame representations in e-learning, which is devoted to theoretical study of frame models and analysis of their known applications.

Accumulative frame models (AFMs) are relatively autonomous, logically distinct units, that are distinguished by the possibility of multiple use in different situations. For their description, a named frame structure, consisting of slots is used. The report examines the structure of the frames and the ways in which they are presented, focusing on the tabular, which is best suited for presenting knowledge in the field of programming. The main groups of tasks related to the use of AFM in programming training and their implementation are listed. An example of solving tasks for modifying unknown programming code by changing the slots' contents of frames, introducing new and independent design and programming code, that has been experimented in the practical classes of students of specialty "Automation and Computer Systems" in the course "Algorithms and Data Structures" at Faculty of Technics and technologies - Yambol.

The use of frames as a mean of presenting knowledge in artificial intelligence has been suggested by M. Minsky in 1974. Frames are used in the theory and practice of e-learning (in Physics, Mathematics, English, etc.) as a mean for classifying and cataloguing of learning units in the text. The relevant methodology is based on the students' solution of the task "reading with comprehension of the learning text", coupled with frames creation. The paper introduce the term "accumulative frame model". The advantages of the proposed model are the possibility of its non-standard applications in e-learning such as automated generation of metadata and test items to learning materials, segmentation of text to learning units, data extraction of text, etc. The paper presents stages of a study for development a software system for designing, creating and using accumulative frame models in the e-learning.

Abstract: The paper explores the possibility for automated accumulation and aggregation of data for quality evaluation of higher education. A model for automated accumulation and aggregation of data specified for institutional accreditation by the NEAA criteria system is proposed. An analysis of the possibilities for automated data accumulation for evaluation of the criteria from the Criteria system for institutional accreditation of higher schools of NEAA has been made. Experiments are performed for data accumulation from university information systems.

A relation between the terms frame and programming invariant is established on the basis of the term Accumulative frame pattern (AFP). As a consequence two methods for initial training in programming are proposed and experimented. The first method (" source code reading and understanding ") refers to training for analysing unknown program code by detecting programming in-variants and identifying their parameters. The second method (" program design and implementation ") is limited to: initial presentation of the solution as a number AFPs of programming invari-ants; modification (if necessary) and valuation of their parameters; implementation of a program code as a set of program frames (possibly modified) of the respective AFP; testing and correction. As a conclusion some other prospective directions and possibilities for using AFP are also commented , including the creation of a Virtual Laboratory for Training in Programming. Keywords: training in programming, reading and understanding, knowledge representation by frames, virtual laboratory 1. Увод Фреймът се използва за представяне на минималната структурирана информация, която дефинира еднотипни явления, събития, ситуации или процеси (Korzh, 2016). Понятието " фрейм " е въведено в изкуствения интелект от М. Мински (Minsky M., 1974) като средство за представяне на знания. Под формата на т. нар. " фреймови опори " успешно e ползвано в обучението по математика, физика, химия, история, английски език и др. (Gurina et al. 2007). В случая на обучение за придобиване на умения и стил по програмиране, фреймови модели са предложени в (Somova et al., 2014) под името " инварианти " (пасажи от програмен код, които с малки неструктурни промени се срещат в решенията на задачи за програмиране). На базата на понятието " акумулативен фреймов модел " – АФМ (Panayotova et al., 2017), в работата е направена връзка между понятията " инвариант в програмирането " и " фрейм " , предложени и експериментирани са методики, подходящи за провеждане на начално обучение по програмиране. 2. Фреймови модели в обучението по програмиране Инвариантите в програмен текст могат да се разглеждат като относително самостоятелни, логически обособени единици, които се отличават с възможността за многократно използване в различни ситуации. Същото се отнася и за АФМ, който може да се представи като именована фреймова структура, съставена от слотове. От своя страна,