Available online at www.sciencedirect.com ScienceDirect Procedia Computer Science 104 (2017) 152 – 159 ICTE 2016, December 2016, Riga, Latvia Personalized Planning of Study Course Structure Using Concept Maps and Their Analysis Raita Rollandea,*, Janis Grundspenkisb Engineering Research Institute “Ventspils International Radio Astronomy Centre” of Ventspils University College, Inzenieru Street 101a, Ventspils, LV-3601, Latvia b Riga Technical University, Faculty of Computer Science and Information Technology, Department of Artificial Intelligence and Systems Engineering, Daugavgrivas Street 2–233, Riga, LV-1007, Latvia Abstract Four graphs for personalized study planning constitute the personalized study planning framework, namely, a graph representing a conceptual structure of study program; a graph representing study course; a graph visualizing each topic of study course using concept map; a graph representing learning objects This paper deals with the third graph – a graph displaying study course topic structure and knowledge assessment Authors describe concept map based knowledge evaluation system integration possibilities with personalized study planning prototype and usage in personal study planning In order to perform the structure analysis of the concept maps authors propose to use of the methods of structure analysis to calculate the ranks for the nodes of the graphs thus detecting the most significant nodes in the graph structure The calculation of ranks for the graph nodes allows detecting the most essential concepts in the concept map © Published by Elsevier B.V.B.V This is an open access article under the CC BY-NC-ND license ©2017 2016The TheAuthors Authors Published by Elsevier (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of organizing committee of the scientific committee of the international conference; ICTE 2016 Peer-review under responsibility of organizing committee of the scientific committee of the international conference; ICTE 2016 Keywords: Personalized education; Graphs; Concept maps; Knowledge assessment; Structural modelling and analysis; Introduction In this chapter the personalized study planning framework is described based on the set of graphs and elaborated in previous research1 * Corresponding author Tel.: +371 63629657; fax: +371 63629660 E-mail address: venta@venta.lv 1877-0509 © 2017 The Authors Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of organizing committee of the scientific committee of the international conference; ICTE 2016 doi:10.1016/j.procs.2017.01.093 Raita Rollande and Janis Grundspenkis / Procedia Computer Science 104 (2017) 152 – 159 To accomplish the personalized study planning starting from the creation of study programme plan and ending with the choice of learning objects, there is developed the framework of personalized study planning based on the following set of graphs (see Fig 1)2,3,4: x Representing a conceptual structure of study program G1(V1,Q1) allows to design individual study plan x Representing study course G2(V2,Q2) allows to develop individual learning scenario x Visualizing each topic using concept map G3(V3,Q3) ensures mapping of each topic to the corresponding concept map x Representing learning objects G4(V4,Q4) describes each concept with learning object The graphs of personalized study planning framework are mutually related At first, the learner in order to create an individual study plan needs to select courses to be included in the individual study plan (graph G1) After that the learner may choose the study courses which he/she wants to master Study course structure is described using graph G2 and that allows to develop individual learning scenario In the study course structure graph G2 learner needs to choose course topic and graph G3 describes at the next level the concepts of each topic and their mutual relationship The learning objects of topics and concepts, that are available for acquiring knowledge are described in the fourth graph G4 of the framework of personalized study planning Fig General structure of personalized study planning framework This framework allows any learner to tailor a desired study program by adapting the modularized curriculum structure and to choose the suitable learning strategy for each study course5,6 In order to fulfil a personalized study planning framework, a Study Planning System (SPS) as a prototype4 has been developed according to the personalized study planning framework Based on the graph G2(V2, Q2) this prototype allows individually to visualize and design the study plan according to the graph G1(V1, Q1) as well as the choice of a study course Input of learning objects is provided if the graph for representing learning objects G4(V4, Q4) is not visually represented in the prototype Study Planning System prototype does not implement the graph for presenting course topics with a concept map G3(V3, Q3) In the given paper the authors deal with the graph G3(V3, Q3) integration possibilities with concept-map-based knowledge assessment system IKAS and the concept map structure analysis The following chapter describes the graph G3 in more details Concept maps of the course topics In course structure graph G2 learner needs to choose course topic as shown in Fig and at the next level graph G3 describes the concepts7 of each topic and their mutual relationships Concept maps were introduced in 1972 by J D Novak8 The presentation of knowledge in the form of a graph is called a concept map9 Graph G3 (V3, Q3) is used for representing the concepts and their relationships The concepts are represented by a set of un-empty nodes V3, and the relations between concepts are represented by a set of un-empty links Q3 The relation between concepts is described by the semantics of the links A concept can describe an event or an object Two or more correlated concepts form a sentence8 Cross links are another concept map characteristics These are relationships between concepts from different domains Cross-links allow us to see in what way some domains of knowledge represented on the map are related to each other8 153 154 Raita Rollande and Janis Grundspenkis / Procedia Computer Science 104 (2017) 152 – 159 Concept maps can be helpful in the study process and when evaluating knowledge8,10 Because they laconic reflect all acquired concepts to learner Concept maps can be displayed in several levels: small concept map for displaying a single study course, a large concept map for displaying the whole study program With the help of the concept maps it is possible to demonstrate the concept cross links The hierarchical arrangement of the concept maps demonstrates the learning sequence of the concepts11 The learner perceives the training as “conceptually more transparent”8 if concept maps are used for showing the structure of the course topic Not all learners identify successfully the most significant concepts, but a concept map helps to acquire them Concept maps reflect the structure of learner knowledge The learners may find some study courses that include remembering lots of facts, dates, names, equations etc boring, while the use of concept maps allows them to control the course and acquire it successfully8 Various displaying types of concept maps represent domain complexity12 For displaying simple structures various types of concept maps are used, for example linear, tree, cycle and the stars type of concept map However, for complex structure displaying network type concept maps are used10 To draw a concept map, the authors of the paper8 have identified the following steps: x Domain selection corresponds to identification of the issues to which answers should be provided with the help of concept maps, for example, in relation to the course "Object-oriented modeling" teacher can ask questions "What is the class?" "What are the components of the class?" etc x Identification of the key concepts that apply to this domain is executed when the teacher chooses the main concepts from the course that leads to the completion of the course objectives by that identifying all the key concepts that are answers to the questions raised in previous step x Construction of a preliminary concept map is carried out by moving the concepts together with linking statements, which explains the relationship between the two concepts x Revision and improvement of created concept maps x Determination of cross links The graph G3 can be directed or undirected if shown as a concept map Fig depicts an example of concept map of the graph’s G2 topic “Components of class diagrams” Learners can personalize the way of acquiring knowledge about the specific concept, if the concepts of the graph G3(V3, Q3) are connected with the topics of the course structure graph G2(V2, Q2) The integration opportunities for personalized study planning system and IKAS The content of topics, each of which is presented as a set of concepts in its turn is presented by the personalized study planning system framework containing graph G3 But that is not realized in the personalized study planning system prototype SPS Consequently, the concept-map-based knowledge assessment system IKAS has been selected in order to fully realize the individualized study planning framework Since 2005 IKAS is being developed by Riga Technical University Department of Artificial Intelligence and Systems Engineering13 The process of IKAS development and improvement will continue also in the future9 In personalized study planning system prototype SPS the learners establish their individual study plan (according to the graph G1), then they choose courses from created study plans to start training and open course structure (according to the graph G2) From the course structure the learners choose the course topic The course topic learning steps are possible in two ways First, selecting and acquiring learning objects (according to the graph G4) or then constructing a concept maps (according to the graph G3) using IKAS, as well as the learner's knowledge assessment If the learner after the graph G2 has chosen to continue to work with the graph G4, then, to carry out knowledge test, he or she must return to IKAS which supports graph G3 IKAS is knowledge assessment system that assesses the learner's knowledge of the study subjects It compares teacher-made concept map with concept map constructed by the learner The knowledge assessment system has two groups of users –learners and teachers Teachers must define their concept map first before learners can create their own concept map The IKAS and its environment are described in14 Raita Rollande and Janis Grundspenkis / Procedia Computer Science 104 (2017) 152 – 159 Fig Components of class diagrams The teacher's access mode to create a concept map of the topic "Class diagram components" is shown in Fig Concept map consists of a total 35 concepts Initially defined concepts based on the concept of map structure analysis are five: "Package", "Link", "Class", "Types of links", and "Multiplicity" Other concepts which are chosen by the trainees should be placed in by themselves Fig shows the window of the student’s interface for starting the completion of a concept map for topic "Class diagram components" created in IKAS system’s learner access mode Initially, in the concept map creation window learner see concepts which has been initially defined by a teacher The selection of the concepts is done on the basis of the concept map structure analysis Other concepts are visible at the bottom of concept map creating window The learner creates the concept map with shown in the bellow concepts Then the learner connects them To understand the concept, the learner can open help window The help window displays the concept definition, description and example To connect the concepts the IKAS program window offers types of linking phrases The learner must choose the most appropriate linking phrase between the two concepts Once the trainee has completed a concept map, he/she activates the button , after which IKAS compares the concept map defined by the teacher with the learner's developed concept map and after the button activation detailed learner concept map assessment is eliminated After acquiring the course topic the teacher provides input mark in the course structure graph (G2) and flags that a course topic has been acquired and selects the next topic for concept map construction The personalized study planning system prototype SPS does not include a learner’s knowledge test, while IKAS does In such a way the learner can develop his/her own study plan in the personalized study planning system, determine the sequence of topic acquisition, then select concepts by using IKAS, move to SPS in order to acquire the learning objects and then come back to IKAS in order to carry out a knowledge test 155 156 Raita Rollande and Janis Grundspenkis / Procedia Computer Science 104 (2017) 152 – 159 Fig IKAS teacher access mode to create a concept map of the topic "Class diagram components" Fig Concept map creating window for topic "Class diagram components" After the knowledge test, the learner marks that the topic is acquired in the personalized study planning system, and starts acquiring the next topic The learner continues the task sequence to acquire all the other study course topics until the whole study plan has been accomplished Concept map structure analysis To analyze structure of concept map, the authors have chosen to use offered in book15 approach of structure analysis – structural modelling because it allows to define the significance of elements which is the main difference from other methods Structural modelling is the way of topological modelling based on computerized construction 157 Raita Rollande and Janis Grundspenkis / Procedia Computer Science 104 (2017) 152 – 159 and analysis of models, development of knowledge basis, and the use reasoning procedure15 It started to develop in Riga Technical University at mid 1970s, it is suitable for technical system with physically multiple elements for mathematical modelling in the circumstances of incomplete information15 In structural modelling there are investigated relationships between structure elements, the importance of elements in functioning of the system as well as the assessment of consequences in case of element elimination15 Structural modelling has several methods that allow to perform structure analysis, judge about the role of elements in the structure, and the common characteristics of the structure One of the methods - the calculation of the ranks of graph nodes The acquisition of the ranks allows to define the local degree of importance of the node of whole graph The higher is the rank of the element the closer this element is related to other elements in the structure and the more serious consequences may arise if it is excluded from the structure15 In structural modelling it is distinguished between qualitative and quantitative structure analysis15 Qualitative analysis defines the importance of the graph nodes Quantitative analysis uses the distance between the elements defined in the graph theory To define the importance of graph nodes in the elaborated framework graphs and calculate from G3 the most important concepts, the authors use the qualitative analysis of the structure In order to make qualitative analysis of the structure, ranks should be calculated for the nodes15 There are chosen three methods for rank calculations: x By the local degree of the node R1LP To determine the rank of the elements of the local degree, element input and output nodes have to be defined, and after that the sum of input and output nodes is calculated by which the elements are ranked The highest rank is allocated to the nodes with the highest local degree x By the number of routes in the graph which contain the given node R1CE In this case wider analysis of node mutual relationship is carried out, stating in how many different routes the node is included In order to that, first is stated the set of all routes between input and output nodes After that the number of routes containing the given node is calculated The obtained number of routs for each node is divided with total number of routes found The highest rank is given to the nodes with the greatest value which is obtained dividing the number of routes for each node to the total number of routes in the graph The rank R1CE shows the structural importance of the node x By the number of reachability nodes R2CE In this case there are taken into account all routes that make output from node, but the routes making input into the node are not taken into account In order to find the rank according to the number of achievable nodes, in the matrix line of reachability all the elements are added and the reachability component is obtained, which afterward is divided to total number of nodes in the graph15 The greater is the value, the higher is the rank To calculate the element structural significance, firstly the summary ranking Rsum is calculated: ଶ ଵ ܴ௦௨ ሺ݅ሻ ൌ ܴ ܴா (1) Secondly, the summary ranks are arranged by places and thus a total rank of elements Rtot is obtained Then the element’s structural significance N(i) can be calculated: N (i ) Rtot , Rmax (2) where N(i) – element’s structural significance, Rtot– element’s total rank, Rmax – the maximum value of the sum of rank To calculate the node structural significance usually there are used two ranks: by the number of routes in the graph and by the number of reachability nodes15, and the authors offer for calculation of the node structural significance also to include the third rank which is calculated by the local degree of the node If the nodes are ranked according to their local degrees, then there are analyzed direct links, but indirect links that are essential in 158 Raita Rollande and Janis Grundspenkis / Procedia Computer Science 104 (2017) 152 – 159 complicated systems, are ignored15 Thus the local degree analysis in complicated systems is not actual However, it is different in personalized study planning system where calculation of ranking by the local degrees of nodes is essential when it is necessary to analyze the local information of each study course, topic, or concept separately To calculate the values of node structural significance which is based on rank values, first is calculated the summary rank Rsum1 taking into account ranks ଵ ଶ ଵ ܴ௦௨ ሺ݅ሻଵ ൌ ܴா ܴா ܴ (3) After that to calculate the values of nodes structural significance N(i)1, that is based on rank value, the following formula is used: N (i )1 1 Rtot Rmax (4) For the analysis of the concept map structure there are used concepts of the topic ‘Class Diagram Components’ Reviewing the results of the concept map analyze, it can be concluded: x Performing the structure analysis by local degrees the highest rank is calculated for the concept ‘Types of Links’, that means that this concept has most direct links with other concepts The next significant concepts by local degrees are with equal values ‘Class’ and ‘Cardinality’, followed by ‘Visibility’ and with equal ranking values ‘Links’, ‘Operations’ and ‘Attributes’ x Calculating the ranks by the number of routes which contain the given concept, the highest ranking values are for the concepts ‘Packages’, ‘Links’, ‘Class’, ‘Types of Links’, and ‘Operations’ x Analyzing the structure of the study course topic graph by the number of reachability nodes, most routes are possible from the concepts ‘Packages’, ‘Links’, ‘Types of Links’, ‘Class’, and ‘Cardinality’ Evaluating the results by the structural significance, the first most significant concepts are ranked as follows: ‘Packages’, ‘Links’, then with equal values are ‘Types of Classes’ and ‘Class’, followed by ‘Cardinality’ These are five the most significant concepts In order the learner can properly master the knowledge on topic ‘Class Diagram Components’, it is compulsory to master the most significant concepts in the structure Assessing the knowledge of the learner when comparing the concept map created by the learner to the concept map defined by the lecturer, it is advisable to take into account the significance of the nodes, that means, that more points in the assessment are given if the learner has correctly identified significant concept and its correlation to other concepts than less significant concept Determination of the most important concepts helps the lecturer in knowledge evaluation when creating the concept map While comparing the concept map created by the learner with the concept map defined by the lecturer it is advisable to take into account the importance of the nodes, i.e that more points should be added to the evaluation if the learner has correctly identified the more significant concept and its relations to other concepts than the less significant concept Conclusion The integration of personalized study planning framework graph G3 allows to present the content of study course topics It uses concept maps based knowledge evaluation system IKAS, that allows to acquire study courses and assess knowledge The authors describe integration opportunities of concept map based knowledge evaluation system with personalized study planning prototype and its usage in personal study planning Such integration approach allows to extend functionality of both systems: personalized study planning prototype SPS and concept map based knowledge evaluation system IKAS IKAS can be part of other system architecture The integration results prove that assumption Personalized study planning framework has modular structure, which facilitates integration Further integration of research and extension functionality of personalized study planning prototype SPS Raita Rollande and Janis Grundspenkis / Procedia Computer Science 104 (2017) 152 – 159 can be implemented using other study planning and organizing systems such as Moodle At the same time the authors describe the use of structural modelling methods to analyze personalized study planning structure based on concept map To carry out the qualitative analysis of the concept map structure Ranks are calculated for the concept map structure nodes References 10 11 12 13 14 15 Rollande R Research and Implementation of Personalized Study Planning as a Component of Pedagogical Module Doctoral Thesis Riga: RTU; 2015 195 Grundspenkis J, Rollande R Graph based framework for personalization of education process realized by the tutoring module of intelligent tutoring system In: Proceedings of International Conference „Perspectives in Business Informatics Research”, 2nd International Workshop on Intelligent Educational Systems and Technology-enhanced Learning (INTEL-EDU 2011) Riga, Latvia; 2011 p 216 – 225 Rollande R, Grundspenkis J Representation of study program as a part of graph based framework for tutoring module of intelligent tutoring system In: Proceedings of Second International Conference on Digital Information Processing and Communications (ICDIPC 2012) Klaipeda, Lithuania; 2012 p 108–113 Rollande R, Grundspenkis J Graph based framework and its implemented prototype for personalized study planning In: The Second International Conference on E-Learning and E-Technologies in Education (ICEEE2013) Lodz, Poland; 2013 p 137–142 Rollande R, Grundspenkis J, Mislevics A New Approach of Using Structural Modelling for Personalized Study Planning International Journal of Advanced Computer Science and Applications (IJACSA) Special Issue on Extended Papers from Science and Information Conference; 2014 p 104-113 Rollande R, Grundspenkis J, Mislevics A The use of structural modelling methods for analysis of personalized study planning In: IEEE Technically Co-Sponsored Science and Information Conference London, UK; 2014 p 921–926 Novak JD Learning, Creating, and Using Knowledge: Concept Maps as Facilitative Tools in Schools and Corporations Mahwah, NJ: Lawrence Erlbaum Associates; 1998 272 Novak JD, Canas AJ The Theory Underlying Concept Maps and How to Construct and Use Them Technical Report IHMC CmapTools 2006-01 Rev 01-2008, Florida; 2008 p 1-36 Grundspenkis J Usage Experience and Student Feedback Driven Extension of Functionality of Concept Map Based Intelligent Knowledge Assessment System Communication and Cognition Vol 43 (1-2); 2010 p 1-20 Vanides J, Yin Y, Tomita M, Ruiz-Primo MA Using concept maps in the science classroom Science Scope Vol 28 (8); 2005 p 2731 Novak JD Learning, creating, and using knowledge: Concept maps as facilitative tools in schools and corporations Journal of eLearning and Knowledge Society Vol (3); 2010 p 21 – 30 Meagher T Looking Inside a Student’s Mind: Can an Analysis of Student Concept Maps Measure Changes in Environmental Literacy? Electronic Journal of Science Education Vol 13 (1); 2009 Grundspenkis J, Anohina-Naumeca A Evolution of the Concept Map Based Adaptive Knowledge Assessment System: Implementation and Evaluation Results Scientific Journal of RTU series Datorzinatne Vol 38; 2009 p 13-24 Grundspenkis J MIPITS and IKAS – Two Steps towards Truly Intelligent Tutoring System Based on Integration of Knowledge Management and Multiagent Techniques In: Proceedings of International Conference on e-Learning and the Knowledge Society (eLearning'10); 2010 Osis J, Grundspenkis J, Markovics Z Topological Modeling of Complex Heterogeneous Systems: Theory and Applications Riga: RTU; 2012 407 Raita Rollande, Assistant Professor of the Faculty of Information Technology of Ventspils University College, researcher of the Engineering Research Institute “Ventspils International Radio Astronomy Centre” of Ventspils University College In 2015, Raita successfully defended her Doctoral Thesis at the Riga Technical University Her research interests include intelligent tutoring systems, personalized education, graphs, structural modelling, element ranking and structural analysis Contact her at raita.rollande@venta.lv Janis Grundspenkis, Professor of Systems Theory at Riga Technical University, Latvia In 1993 he received the degree of Habilitated Doctor of Engineering Sciences from Riga Technical University His research interests include development of methods and tools for modelling, analysis, and diagnosis of complex heterogeneous technical systems and development of intelligent tutoring and knowledge assessment systems Contact him at janis.grundspenkis@rtu.lv 159 ... qualitative analysis of the concept map structure Ranks are calculated for the concept map structure nodes References 10 11 12 13 14 15 Rollande R Research and Implementation of Personalized Study Planning. .. topics of the course structure graph G2(V2, Q2) The integration opportunities for personalized study planning system and IKAS The content of topics, each of which is presented as a set of concepts... and the concept map structure analysis The following chapter describes the graph G3 in more details Concept maps of the course topics In course structure graph G2 learner needs to choose course