Halil Ibrahim Bulbul

Determination of Vocational Fields with Machine Learning Algorithm

The importance of vocational and technical training is growing day by day in parallel to the developing technology. It is inevitable to utilise opportunities presented by information and communication technologies in order to determine vocational fields in vocational and technical training in the most efficient manner. In this respect, it is possible to create a more efficient tool compared to the current methods by utilising machine learning which is an artificial intelligence model in energy applications that predicts events in the future depending on the past experiences. In the current study, a software is developed that ensures that the system learns about the successful and unsuccessful choices made in the past by applying “Naive Bayes” algorithm, which is a machine learning algorithm, to the data collected concerning the individuals who turned out to be successful or unsuccessful in the vocational technical training process in energy applications. In the software developed, it is aimed that the system recommends the most suitable vocational field for the individual by according to the data collected from the individual who is in the occupation selection process in field energy applications.

Comparison of Classification Techniques used in Machine Learning as Applied on Vocational Guidance Data

Recent developments in information systems as well as computerization of business processes by organizations have led to a faster, easier and more accurate data analysis. Data mining and machine learning techniques have been used increasingly in the analysis of data in various fields ranging from medicine to finance, education and energy applications. Machine learning techniques make it possible to deduct meaningful further information from those data processed by data mining. Such meaningful and significant information helps organizations to establish their future policies on a sounder basis, and to gain major advantages in terms of time and cost. This study applies classification algorithms used in data mining and machine learning techniques on those data obtained from individuals during the vocational guidance process, and tries to determine the most appropriate algorithm.

Analysis for Status of the Road Accident Occurance and Determination of the Risk of Accident by Machine Learning in Istanbul

The traffic has been transformed into the difficult structure in points of designing and managing by the reason of increasing number of vehicle. This situation has discovered road accidents problem, influenced public health and country economy and done the studies on solution of the problem. Large calibrated data agglomerations have increased by the reasons of the technological improvements and data storage with low cost. Arising the need of accession to information from this large calibrated data obtained the corner stone of the data mining. In this study, assignment of the most compatible machine learning classification techniques for road accidents estimation by data mining has been intended.

FPGA based parallel connection system of separate voltage sources

In this study, parallel connection of separate voltage sources has been realized using cuk converters. For this aim, output voltages of converters are measured continuously and then fixed to a reference value. The PWM signals which are required to control the converters and to sense the current and voltage signals are generated by an FPGA. Since the FPGA has parallel processing feature, voltage and current data of the system are measured at the same time. Thus, the converters are connected in parallel punctually and load sharing between them are provided in order to eliminate some undesired cases like unbalanced loading. In the system realized, two identical cuk converters are used. As cuk converters can operate both in buck and boost modes, output voltages are fixed to the reference voltage successfully, even though input voltages are low or high. Experimental results show that the system achieved operates with a reasonable efficiency ranging from 81% to 88%.