Adem Karahoca

Information system design for a hospital emergency department: A usability analysis of software prototypes

STUDY OBJECTIVE The purpose of this study is to evaluate the usability of emergency department (ED) software prototypes developed for Tablet personal computers (Tablet PCs) in order to keep electronic health records (EHRs) of patients errorless and accessible through mobile technologies. In order to serve this purpose, two alternative prototypes were developed for Tablet PCs: Mobile Emergency Department Software (MEDS) and Mobile Emergency Department Software Iconic (MEDSI) among which the user might choose the more appropriate one for ED operations based on a usability analysis involving the target users. METHODS The study is based on a case study of 32 potential users of our prototypes at the ED of Kadikoy-AHG in Istanbul, Turkey. We examined usability of the prototypes for medical information systems by means of Nielsens heuristic evaluation and cognitive walkthrough methods relying on 7-point scales, and scenario completion success rate and average scenario completion time, respectively. RESULTS The implementation of MEDSI in our case study confirmed the view that the usability evaluation results of iconic GUIs were better than those of non-iconic GUIs in terms of Nielsens heuristic evaluation, effectiveness and user satisfaction. For the whole sample, paired t-test scores indicated that there was a significant difference (p<0.01) between mean values of Nielsens usability scores toward MEDS and MEDSI indicating that MEDSI was evaluated more favorably than MEDS. As for effectiveness of the prototypes, significant differences (p<0.01) were noted between MEDS and MEDSI in terms of both overall scenario completion success rate and average scenario completion time. Similarly, for the full sample of users independent sample t-test scores indicated that MEDSI was perceived significantly more favorable (p<0.01) than MEDS in terms of overall user satisfaction. CONCLUSION The study provides two important contributions to the extant literature. First, it addresses a topic and methodology that serves potentially interesting to the biomedical informatics community. Drawing on good background information and appropriate context, it involves various aspects of usability testing. Another contribution of the study lies in its examination of two different prototypes during the design phase involving the target users.

E-learning success model for instructors’ satisfactions in perspective of interaction and usability outcomes

Abstract E-learning approaches could be handled in a system design view in which the system components and factors have critical roles in order to assure success of whole system. In such an e-learning design view, online instructors (or faculties) have the most critical role as the most important actor. Therefore there is an emerging need for investigating the factors affecting instructors’ performance in e-learning systems. Satisfaction is one of these factors that affect usability of the system which also directly affect instructors’ performance. In this study, factors related to instructors’ satisfaction in e-learning systems have been investigated in order to develop a basic model called “E-Learning Success Model for Instructors’ Satisfactions” which is related to social, intellectual and technical interactions of instructors in whole e-learning system. “E-Learning Success Model for Instructors’ Satisfactions” could be a basic guide for e-learning designers, online instructors and policy makers to understand interaction and usability outcomes related to satisfaction of instructors.

Neural network based motor bearing fault detection

Bearing faults are the biggest single cause of motor failures. The bearing defects induce vibration resulting in the modulation of the stator current. The stator current can be analyzed via wavelet packet decomposition to detect bearing defects. This method enables the analysis of frequency bands that can accommodate the rotational speed dependence of the bearing defect frequencies. In this study, radial basis function neural networks are used to improve bearing fault detection procedure.

Is SMS still alive for education: Analysis of educational potentials of SMS technology?

Abstract Mobile phone, especially SMS concept is a fairly old technology. In spite of the fact that there are some applications of using SMS capabilities of mobile phones in education, the usage is not widely in all educational areas. Still, there is a need to understand how SMS technology could be applicable for education in different ways. Hence, in this study, authors’ investigate possible educational use of SMS by providing a detailed analysis of the technology and examples of different research studies of successful implementations in education. This study may be a guide for school administrators and policy makers to understand educational technology potential of SMS by providing benefits and drawbacks of the technology.

Predicting existence of Mycobacterium tuberculosis on patients using data mining approaches

Abstract A correct diagnosis of tuberculosis (TB) can be only stated by applying a medical test to patient’s phlegm. The result of this test is obtained after a time period of about 45 days. The purpose of this study is to develop a data mining(DM) solution which makes diagnosis of tuberculosis as accurate as possible and helps deciding if it is reasonable to start tuberculosis treatment on suspected patients without waiting the exact medical test results or not. In this research, we proposed the use of Sugeno-type “adaptive-network-based fuzzy inference system” (ANFIS) to predict the existence of mycobacterium tuberculosis. 667 different patient records which are obtained from a clinic are used in the entire process of this research. Each of the patient records consist of 30 separate input parameters. ANFIS model is generated by using 500 of those records. We also implemented a multilayer perceptron and PART model using the same data set. The ANFIS model classifies the instances with an RMSE of 18% whereas Multilayer Perceptron does the same classification with an RMSE of % 19 and PART algorithm with an RMSE of % 20. ANFIS is an accurate and reliable method when compared with Multilayer Perceptron and PART algorithms for classification of tuberculosis patients. This study has contribution on forecasting patients before the medical tests.

Robotics teaching in primary school education by project based learning for supporting science and technology courses

Abstract This study investigates robotics education to support science and technology courses for primary school education in Private Evrim College in Istanbul. This research is a kind of case study on a group of students who are having a robotics education. The robotics education is not only providing information to students about robot making also help them to improve their skills and abilities. First starting point is introducing equipments for designing electronic circuits and teaching fundamental principles for robot making. Robotics education performed with 16 students between 10 and 15 years old and they are distributed in 4 groups. In the research, it will be seen that students can learn and apply their skills on designing circuits while they are discussing with classmates by the means of collaborative learning and project based learning. Also, students have been attended to the race in their schools. Furthermore the winner group is competed in a robotics competition which is held every year in Istanbul Technical University. Consequently, the findings and evaluations help to understand that robotics education to support students’ life, moreover, affecting their science performance and relationships with friends in class positively.

Simultaneous feature selection and ant colony clustering

Clustering is a widely studied problem in data mining. Ai techniques, evolutionary techniques and optimization techniques are applied to this field. In this study, a novel hybrid modeling approach proposed for clustering and feature selection. Ant colony clustering technique is used to segment breast cancer data set. To remove irrelevant or redundant features from data set for clustering Sequential Backward Search feature selection technique is applied. Feature selection and clustering algorithms are incorporated as a Wrapper. The results show that, the accuracy of the FS-ACO clustering approach is better than the filter approaches.

Dosage planning for diabetes patients using data mining methods

Abstract The condition of high blood sugar (glucose) level is called as diabetes mellitus. Cause of this disease can be either insufficient insulin production or improper response of body cell to insulin. Diabetes patients should use different drugs in order to keep their blood sugar level within normal range values. The purpose of this study is to develop a data mining model for which will predict a suitable dosage planning for diabetes patients. Medical records of 89 different patient records were used in this study. 318 diabetes assays were extracted using these patient records. ANFIS and Rough Set methods were used for dosage planning objective. According to the results of ANFIS and Rough Set methods, ANFIS is a more successful and reliable method for diabetes drug planning objective when compared to Rough Set method.

Prediction of cyclosporine A blood levels: an application of the adaptive-network-based fuzzy inference system (ANFIS) in assisting drug therapy.

ObjectiveTherapeutic drug monitoring (TDM) is a procedure in which the levels of drugs are assayed in various body fluids with the aim of individualizing the dose of critical drugs, such as cyclosporine A. Cyclosporine A assays are performed in blood.MethodsWe proposed the use of the Takagi and Sugeno-type “adaptive-network-based fuzzy inference system” (ANFIS) to predict the concentration of cyclosporine A in blood samples taken from renal transplantation patients. We implemented the ANFIS model using TDM data collected from 138 patients and 20 input parameters. Input parameters for the model consisted of concurrent use of drugs, blood levels, sampling time, age, gender, and dosing intervals.ResultsFuzzy modeling produced eight rules. The developed ANFIS model exhibited a root mean square error (RMSE) of 0.045 with respect to the training data and an error of 0.057 with respect to the checking data in the MATLAB environment.ConclusionANFIS can effectively assist physicians in choosing best therapeutic drug dose in the clinical setting.

Diagnosis of diabetes by using adaptive neuro fuzzy inference systems

Most of discoveries indicate that the best way to overcome diabetes is to prevent the risks of diabetes before becoming a diabetic. With this opinion, we would like to find a way to estimate diabetes risk, according to some variables such as age, total cholesterol, gender or shape of the body. Due to having fuzzy input and output (glucose rate) values and because of that dependent variable have more than 2 values (unlike binary logic), ANFIS and Multinomial Logistic Regression should be executed for comparison. Then the results were benchmarked. As a result, in case of that there is a system which contains fuzzy inputs and output, ANFIS gives better results than Multinomial Logistic Regression for diabetes diagnosis.