A common method to boost the efficiency of instance-based classifiers, while maintaining high accuracy, is to decrease the training set size by substituting it with a smaller, representative subset. This is typically achieved by utilizing data reduction techniques. The latter enjoy wide applicability in handling single-label classification tasks. This is not the case though in cases of multi-label data, where each instance may be associated with not just one, but a number of classes. In the present paper, we adapt a popular single-label data reduction technique - the Edited Nearest Neighbor (eNN) rule - to handle multi-label data. In single-label classification, eNN focuses on the removal of noise and close border instances making the dataset clear and the borders well-separated. The core idea is that instances with a different class than the majority of their neighbors are considered noise and are removed. In the context of multi-label data, label boundaries tend to be ambiguous, and the notion of noise is not clearly defined. Nevertheless, we hypothesize that instances whose labelsets significantly differ from those dominating their local neighborhood can be treated as noise and their removal serves to condense the training set. Based on this principle, we propose three (3) new eNN variations and test them in practice. Experimental tests and statistical analysis conducted across nine (9) diverse multi-label datasets indicate that the proposed algorithms reduce significantly the size of the datasets, without compromising on classification accuracy.

Reducing the size of the training set, that is, replacing it with a condensing set, while maintaining the classification accuracy as much as possible is a very common practice to speed up instance-based classifiers. Data reduction techniques, also known as prototype selection or generation algorithms, can be used to accomplish this. There are numerous such algorithms that can be found in the literature that are effective for single-label classification problems, but the majority of them cannot be used for multi-label data where an instance may belong to multiple classes. Due to the numerous binary condensing sets it creates, the well-known Binary Relevance transformation method cannot be combined with a Data Reduction algorithm. Condensed Nearest Neighbor is a well-known parameter-free single-label prototype selection algorithm. This study proposes three variations of that algorithm for training datasets with multiple labels. An experimental study that we conducted over nine distinct datasets shows that our three proposed approaches provide good reduction rates while not tampering with the classification rates.

Reduction through Homogeneous Clustering (RHC) and its editing variant (ERHC) are effective data reduction techniques for the k-NN classifier. They are based on an iterative k-means clustering task that discovers homogeneous clusters. The centers of the resulting homogeneous clusters constitute the instances of the reduced training set. Although RHC and ERHC are quite fast compared to several well-known data reduction techniques, the iterative execution of k-means clustering renders both of them inappropriate for data reduction tasks that need to be performed quickly, especially, when run over large training datasets. The present paper proposes simple and very fast variations of the algorithms, which are appropriate for such environments. The variations are called RHC2 and ERHC2 and replace the complete execution of k-means clustering with a fast task that assigns instances to the class centers. The experimental study based on fourteen datasets, and, the corresponding statistical tests, show that the proposed RHC2 and ERHC2 variations are very fast and, at the cost of a small penalty on classification accuracy, they achieve higher reduction rates than their predecessors and other two well-known data reduction techniques. They are good candidates when fast reduction on large datasets is required.

The use of online videos is a common practice today in education and this is evident from the abundance of free online courses that are delivered by universities, educational organizations as well as individual professionals. In most cases, the videos that exist in educational platforms are linear without any scope for interactivity and in cases where there is interactivity this is mainly restricted to in-video quizzes. The aim of this paper is to present ways for going beyond the standard in-video quizzes by augmenting videos with various interactive exercises and games that test learners' knowledge. Furthermore, the paper presents a learning environment built using the concepts explained as well as how this environment was used and evaluated in educational settings. The outcome of the evaluation showed that learners attained positive opinions regarding the interactive environment features and its effectiveness in learning.

On line video is a powerful tool for e-learning and this is evident from a number of reports, research papers and university initiatives, which portray that online video is becoming an important medium for delivering educational content. Therefore, research that focuses on how students view educational videos becomes of particular interest and in previous work we argued that in order to efficiently analyze learner viewing behavior we should deploy tools that log the learner activity and assist usage analysis and data mining. Working towards this direction, a framework for recording and analyzing learner behavior was presented together with findings of applying the framework into educational settings. In this paper, we continue this work by presenting a set of metrics that can be derived from the framework and be used to measure learner engagement and video popularity. These metrics in conjunction with the data mining method of clustering are then used to gain insights into learner viewing behavior.

Online video use in education is an expanding trend and research that focuses on how students view educational videos becomes of particular interest. In our previous work, we argued that in order to conduct a detailed analysis of the learner viewing behavior we have to deploy tools that log the learner activity and assist statistical analysis and data mining. Working towards this direction, a framework for recording and analyzing learner behavior was presented. This paper presents findings of applying this framework into educational settings.

Online educational videos are increasingly used by a great number of educators. Besides the typical linear, non-interactive videos, educational videos today can contain other features such as interactivity and quizzes. This paper provides a framework for recording, monitoring and analyzing learner activity behavior while watching or interacting with online educational videos. More specifically it proposes: (a) a software specific methodology for capturing learner activity data, (b) a data model for storing the activity data, and, (c) modules to monitor and visualize learner viewing behavior.

Some of the most commonly used classifiers are based on the retrieval and examination of the k Nearest Neighbors of unclassified instances. However, since the size of datasets can be large, these classifiers are inapplicable when the time-costly sequential search over all instances is used to find the neighbors. The Minimum Distance Classifier is a very fast classification approach but it usually achieves much lower classification accuracy than the k-NN classifier. In this paper, a fast, hybrid and model-free classification algorithm is introduced that combines the Minimum Distance and the k-NN classifiers. The proposed algorithm aims at maximizing the reduction of computational cost, by keeping classification accuracy at a high level. The experimental results illustrate that the proposed approach can be applicable in dynamic, time-constrained environments.

In this paper we present our findings on teaching OOP with BlueJ in the context of a one-semester programming course. We organize our findings, i.e., the difficulties, the errors, and the misconceptions that students encounter, in two categories: (a) difficulties attributed to the special characteristics of OOP, and, (b) difficulties that may be attributed to the features of the programming environment

This paper discusses the pilot testing and evaluation of a database-driven Web-based programming environment called WIPE (Web integrated programming environment). WIPE is a teaching tool for secondary education students that are introduced to the principles of programming. The programming environment was used in secondary schools in Greece and the results of its evaluation demonstrate that it successfully deals with the difficulties novices meet.

This paper presents an overview of the design principles and the evaluation of a new programming environment, WIPE (Web Integrated Programming Environment), designed specifically to teach novices the fundamentals of programming. The environment is designed for use in secondary education as a first programming course, in order to help students become familiar with the main programming concepts.

Here we report on our implementation of a hybrid-indexing scheme (bitmap-tree) that combines the advantages of bitmap indexing and file inversion. The results we obtained are compared to those of the compressed inverted file index. Both storage overhead and query processing efficiency are taken into consideration. The proposed new method is shown to excel in handling queries involving set operations. For general-purpose user queries, the bitmap-tree is shown to perform as good as the compressed inverted file index.

We describe a Web-based programming environment that serves for the teaching of the basic principles of programming. The environment is accessible via the Internet through a Web browser and uses a specially featured compiler to translate the source code of the programming language into a pseudo assembly code. Supplementary programming tools accompany the compiler and a database is used for storing the intermediate remits (i.e., the code students develop while trying to solve an exercise). In addition, there is a complete user, file and group management environment and a tool to automate the testing of the student solutions to programming exercises.

In the last two decades, a large amount of research has been conducted in an effort to form a model of student behavior when they try to solve algorithmic or programming problems. The construction of the model is based on the analysis of many types of data, such as for example: (a) the characteristics of the programming languages the students work with, (b) the strategies of the solution that the students follow, and (c) the characteristics of the proposed problem. However, we must observe that modeling is often not based on long-term observations of actual teaching and the proposed problems are usually quite simple.In this paper we attempt to examine a variety of aspects of students' behavior when they learn to program. More specifically, we study: the strategies students use in order to develop and validate a program; the possible role of students' errors in the development of their programs; and the methods students use to deal with these errors. The study was carried out on 90 second-semester CS students who worked in pairs during the 2-hour lab session. They were given a brief description of the Binary Search algorithm and were asked to implement it using AnimPascal. In this study we present the results we obtained from the analysis of the successive versions of students' programs. Based on these results we propose teaching methods to help students overcome the difficulties they face when they learn programming.

The paper presents a simple programming language, called X, and an educational programming environment, called X-Compiler, designed to introduce students to programming. X-Compiler can be used to edit, compile, debug and run programs written in X, a subset of Pascal. X-Compiler could be didactically interesting because of the following features: (a) users can watch the intermediate steps of the execution of a program: source code compilation, correspondence of source and pseudo-assembly code during execution, register content, and intermediate values of user and temporary system variables; also, they can edit the produced pseudo-assembly code and re-execute it, (b) there are many detailed and explanatory messages that can guide novice programmers when debugging their programs and, in general, help them write better programs.

This paper describes an educational programming environment, called AnimPascal. AnimPascal is a program animator that incorporates the ability to record problem-solving paths followed by students. The aim of AnimPascal is to help students understand the phases of developing, verifying, debugging, and executing a program. Also, by recording the different versions of student programs, it can help teachers discover student conceptions about programming. In this paper we describe how our system works and present some empirical results concerning student conceptions when trying to solve a problem of algorithmic or programming nature. Finally, we present our plans for further extensions to our software.