Miodrag Ivković

Mobile learning in practice

This paper describes the implementation and evaluation of a system used for mobile learning. It describes how such a system can be monitored, and what must be taken into consideration for the deployment of the system on mobile devices. It is also leaves room for the modification of the system.

Analysis of basketball games using neural networks

Data mining is a technology in data analysis with rising application in sports. Basketball is one of most popular sports. Due to its dynamics, a large number of events happen during a game. Basketball statisticians have task to note as many of these events as possible, in order to provide their analysis. In this paper, we used data from the First B basketball league for men in Serbia, for seasons 2005/06, 2006/07, 2007/08, 2008/09 and 2009/2010. During these five seasons, total of 890 games were played. Data were collected for individual players, so it was necessary to adapt these in order to show statistics for a whole team. These data were analyzed using feedforward technique in neural networks, which is the most often used technique in analyzing nonlinear sports data. As a final result, we concluded that the most important elements in basketball are two-point shots under the hoop and defensive rebound, i.e. game “in paint”.

MRI Versus 131I Whole-Body Scintigraphy for the Detection of Lymph Node Recurrences in Differentiated Thyroid Carcinoma

OBJECTIVE The aim of this study was to compare the clinical usefulness of MRI and radioiodine (¹³¹I) whole-body scintigraphy for the detection of lymph node metastases in differentiated thyroid carcinoma (DTC). MATERIALS AND METHODS After surgery and ¹³¹I therapy, 40 patients with DTC underwent ¹³¹I whole-body scintigraphy and MRI. Each patient was clinically suspected of having or already had evidence of nodal recurrences (confirmed by laboratory studies, cytologic analysis, or whole-body scintigraphy). Planar whole-body scintigraphy was done after administration of 111 MBq of ¹³¹I, and MRI was done using spin-echo T1- and T2-weighted imaging, T1-weighted spin-echo imaging with fat suppression, and STIR sequences. RESULTS MRI detected nodal metastases as partly or entirely cystic and as heterogeneously enhanced on contrast-enhanced T1-weighted images. Hyperintense cystic areas appeared on T1- and T2-weighted images and STIR sequences in 57% of cases. Nodal metastases showed extracapsular spread in 24% of patients. MRI results were true-positive in 76%, true-negative in 90%, false-negative in 24%, and false-positive in 11% of cases, whereas ¹³¹I whole-body scintigraphy results were true-positive in 71%, true-negative in 91%, and false-negative in 29% of cases. There were no false-positive results of ¹³¹I whole-body scintigraphy. False-negative whole-body scintigraphy was induced by tumor dedifferentiation. The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of MRI were 76.2%, 89.5%, 82.5%, 88.9%, and 77.3%, respectively, and the corresponding values for ¹³¹I whole-body scintigraphy were 71.4%, 100%, 85%, 100%, and 76%, respectively. CONCLUSION Whole-body scintigraphy is more specific than MRI in the detection of nodal metastases in patients with DTC. The principal value of MRI is in non-iodine-avid recurrences and in evaluation of mediastinal foci.

Implementation of integrated learning system within youth counseling

This paper describes the functionalities and characteristics of an integrated learning system used in youth counseling. It describes how such a system can be implemented, and what must be taken into consideration for the deployment of the system on mobile devices. It is also presents the results of survey that was conducted in development phases.

Appliance of neural networks in Basketball - Basketball board for Basketball Referees

What is presented in this paper is a solution that bears large applicable values depending on implementation and realization. The paper describes the methods and the developed software which, based on the action and the movement of the ball, using the neural network, determines the movement of a basketball referee on the court, in order to gain the best view of the action. The solution is developed in Microsoft Visual Studio 2010, written in the programming language C#, reffering to AForge .NET Framework for the support in the aspects of configuring, training and usage of neural networks. AForge .NET Framework is published with LGPL v3 licence.

A system for deductive prediction and analysis of movement of basketball referees

This paper presents us with an automatic prediction and analysis of basketball referees movement which is useful for educational software. Such software would be very beneficial in training the young basketball referees. The paper proposes that the movement prediction of basketball referees can be achieved with a multilayered perceptron neural network. Network will reason on the basis of a ball movement during a play action. Proposed neural network will be trained with a modified Back Propagation algorithm which essentially presents a special algorithm for a multiple dependent Time Series prediction. In this paper, we will also describe initial designs of a neural network structure that, we believe, would better suit the nature of a multiple dependent Time Series prediction problems. The aforementioned educational software is capable of determining whether a referee was moving properly in a certain situation or not. Determination is possible on the basis of numerical values that are calculated by simulating the human visual field. The referee’s horizontal field of view simulation is based on the standard set by the American Optometric Association. It is implemented through a modified Sweep and Prune algorithm which is also discussed in this paper.

Optimization of bulk operation performances within Oracle database

The Oracle database enables use of bulk operation in order to reduce time needed for commands execution over a large amount of data. Operations are called using PL/SQL programming language. Advantage of these operations is that they transfer a large amount of data in a single step to SQL engine where they are executed. Bulk operation may also be used with a limitation to maximal number of lines transferred in a single step between PL/SQL engine and SQL engine. This limitation is specified by the LIMIT parameter. In this paper, performances of inquiry are compared to different values of LIMIT parameter. In this way, we have shown which amount of data is optimal for transfer in a single step, and how further increase influences performances. It is also described what influences optimal amount of data that may be transferred in a single step. Different configurations of memory structures were tested within the Oracle itself (SGA and PGA areas) as well as how their changes influence performances. Influence of data types on optimal amount of data was compared, as well as a role of a processor. The method used for verification of optimal values was arithmetical means of execution times for certain inquiries over database. Measurements were done on the same data in several occasions in order to obtain highly precise value and to eliminate possible background processes that are executed in the computer.

Information retrieval for unstructured text documents in Serbian into the crime domain

The reform and modernisation of public sector based on wide application of information-communication technologies (ICT) is considered as one the key elements of futher development of information society in the Republic of Serbia. The trends of development of many e-Government services for many countries in the world, indicate the necessity of application natural language processing - NLP and in e-Government services Republic of Serbia. When performing many of the natural language processing tasks it is needed that all forms of a word with the same meaning has the same form. In general, many documents that are available e-Government services are not structured, so from them is very difficult to isolate some forms (knowledge) that exist in them. The process of finding useful information from such documents in Serbian language, which normally represents highly inflectional language, is one of the key elements of the modern information society in the Republic of Serbia.

Modeling of an ontology of the administrative task

Administrative tasks are a crucial part of the business process in Public Administration. Ontologies, generated from the workflow of administrative processes, can provide significant improvements and reduce the time of modelling, testing and integration in the process of building an information system of the Public Administration (PA). In this paper, the authors propose an approach for building an ontological model of the administrative task (OMAT) of the Public Administration. The annotation of the proposed ontology with elements of user interface components and presented algorithm for a direct transformation of the semantic description of the administrative task in the user interface components, contribute to the improvement of information systems modeling techniques. The proposed approach, with standardized formalism in the method of creating an ontology of the administrative process (task), can be applied to any administrative process.

Using Neural Networks in Preparing and Analysis of Basketball Scouting

Basketball is the one of most popular sports, in Serbia and in the world. It is a team sport. Actors of a basketball game are the players from two opposing teams, their team officials with coaches, assistant coaches, doctors and officials (commissioner, referees, table officials, statisticians). Every team, depending on a league or competition, may have no more than 12 or 10 players per game, and of those 5 are actively engaged in the game [1]. Basketball game would not start without 5 players from each game on the court. In Europe, regular basketball game is divided into 4 quarters of 10 minutes each, while in NBA every quarter is 2 minutes longer. If a result is draw, after regular time additional time of 5 minutes is played, as many times as necessary to decide a winner of a game.