REDUCTION OF DIMENSION FOR PREDICTION OF PROGRESS IN PROBLEMS OF MEDICAL EDUCATION: AN APPROACH BASED
Background. System research in medical science and education is often based on data with a large number of attributes. Therefore, the problem of reducing the dimension of the data is actual, while keeping data with as much variation as possible. At the same time, such problems can be considered as problems of machine learning, which can be solved using artificial intelligence algorithms (for example, artificial neural networks.) The interpretation of the results of the application of classification algorithms without prior reduction of the dimension of such tasks is a difficulty for the decision maker. The results may be simplified by the reduction of the dimension of the task.
Materials and methods. The method proposed and applied in this paper is described in terms of machine learning. Namely, the mathematical description of the tasks of machine learning in research in the field of medical education is based on the following data. We have a set consisting of tuples. Depending on the problem under consideration, certain sets of these tuples will be used for training, testing and prediction. An arbitrary tuple consists of input data (we will call them according to the established terminology of machine learning as attributes) and output data, which are attributes of the class. As a result of reducing the dimension, we obtain a certain numerical matrix, where these data can be used in variety of problems of machine learning. In order to study the possibilities of reducing the dimension of the medical education assessment tasks, the results of the medical students at TSMU were considered (the volume of the general population was 248).
Results. In the work an approach which is based on application of the method PCA with the purpose of determining principle reasons (current and final progresses), which influence on the results of license integrated exam, is offered.
Using the visualization, approach allows us to represent data on a plane in the coordinates of the first two main components. This allows us to identify the main changes in the attributes that cause the data to belong to certain groups. So, using such a visualization to the student data, we can get graphic representations of the «main» performance indicators (current and final) that predict the passing the licensed integrated exam «Step 1».
Conclusions. Reduction of dimension is a very important element in the analysis of statistical data in medical education. Due to reducing the dimensions, we achieve not only that we simply need to analyze less data. The most important is that after the reduction, they often show us more information than before reducing the dimension. Many dependencies in medical education become more readable. It often happens that the amount of data we have is beyond the scope of our perception and then it is very easy to get lost in their analysis.
Martseniuk V. P. (2007). Vprovadzhennia v navchalnyi protses kompiuternykh tekhnolohii [Implementation of the computer technology educational process]. Medychna osvita (Medical education), 2, 40-41. [In Ukrainian].
Martseniuk V. P. (2008) Rozrobka i vprovadzhennia systemy elektronnoho navchannia v Ternopilskomu derzhavnomu medychnomu universyteti imeni I. Ia. Horbachevskoho [Development and introduction of e-learning system at Ternopil State Medical University named after I. Ya. Gorbachevsky]. Medychna osvita (Medical education), 2, 74-75. [In Ukrainian].
Martseniuk V. P., Semenets A. V. (2009). O strukture bazyi dannyih informatsionnoy sistemyi proverki znaniy v meditsinskom obrazovanii [On the structure of the database of information system for testing knowledge in medical education]. Shtuchnyi intelekt (Artificial Intelligence), 1, 267-277. [In Russian].
Martseniuk V. P., Semenets A. V. (2009). Kontseptualnyie podhodyi k strukture informatsionnoy sistemyi proverki znaniy v meditsinskom obrazovanii [Conceptual approaches to the structure of the information system for testing knowledge in medical education]. Kibernetika i vyichislitelnaya tehnika (Cybernetics and Computer Engineering), 156, 18-27. [In Russian].
Kovalchuk L. Ia., Martseniuk V. P. (2008). Vprovadzhennia v navchalnyi protses kompiuternykh tekhnolohii [Implementation of the computer technology educational process]. Medichna іnformatika ta іnzhenerіya (Medical Informatics & Engineering), 1, 14-16. [In Ukrainian].
Kovalchuk L. Ia., Martseniuk V. P., Selskyi P. R. (2012). Obiektyvizatsiia systemy otsiniuvannia znan studentiv u Ternopilskomu derzhavnomu medychnomu universyteti imeni I. Ya. Horbachevskoho [Objectivization of the student assessment system at the I. Ya. Gorbachevsky Ternopil State Medical University]. Naukovyi zhurnal MOZ Ukrainy (Scientific journal of the Ministry of Health of Ukraine), 2, 104-110. [In Ukrainian].
Kovalchuk L. Ia., Martseniuk V. P., Selskyi P. R. (2012). Obgruntuvannia vykorystannia informatsiinykh tekhnolohii dlia pidhotovky likariv simeinoi medytsyny ta pokrashchennia yakosti medychnoi dopomohy na pervynnomu rivni [Substantiation of the use of information technologies for the training of family medicine doctors and improvement of the quality of medical care at the primary level]. Klynycheskaia ynformatyka y Telemedytsyna (Clinical Informatics and Telemedicine), 8 (9), 141-145. [In Ukrainian].
Martseniuk V. P. (2015). Informatsiina systema upravlinnia yakistiu pidhotovky fakhivtsiv u vyshchii medychnii osviti [Information system of quality management of specialists training in higher medical education]. Ternopil: TDMU. [In Ukrainian].
Martseniuk V. P., Stakhanska O. O. (2010) Analiz rezultativ semestrovykh kompleksnykh testovykh ispytiv v medychnii osviti na osnovi koreliatsiinykh pokaznykiv uspishnosti ta bahatoparametrychnoi neiromerezhevoi klasteryzatsii [Analysis of the results of semester complex test exams in medical education on the basis of the correlation rates of success and multiparameter neural network clusterization]. Medichna mformatika ta mzheneriya (Medical Informatics & Engineering), 1, 53-57. [In Ukrainian].
Martseniuk V. P., Semenets A. V., Stakhanska O. O. (2010). Neiromerezheve prohnozuvannia skladannia studentamy-medykamy litsenziinoho intehrovanoho ispytu «Krok 1» na osnovi rezultativ potochnoi uspishnosti ta semestrovoho kompleksnoho testovoho ispytu [Neural network forecasting of the doctoral students of the licensed integrated exam «Step 1» on the basis of the results of the current progress and the semester complex test test]. Medichna mformatika ta mzheneriya (Medical Informatics & Engineering), 2, 57-62. [In Ukrainian].
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Journal Medical Informatics and Engineering allows the author(s) to hold the copyright without registration
Users can use, reuse and build upon the material published in the journal but only for non-commercial purposes
The majority of Medical Informatics and Engineering Open Access journals publish open access articles under the terms of the Creative Commons Attribution (CC BY) License which permits use, distribution and reproduction in any medium, provided the original work is properly cited. The remaining journals offer a choice of licenses.