Background and aim: Recently, the use of data science techniques in healthcare has been increased remarkably. Community detection as one the important methods of data science is utilized in the health domain. Methods: This paper detects disease areas based on combination of big data and graph mining methods on drug prescriptions. At first, network of prescription is designed, and Louvain algorithm is applied for community detection of 50000 Iranian prescriptions in 2014 gathered from the Iranian Health Insurance Organization. We use modularity metric for validation of the results and the experts’ opinion as the external validation of communities. Results: The outputs are consist of six communities. These communities are labeled based on experts’ opinion that present the disease fields. Conclusion: The Louvain algorithm has the ability to detect the major communities of the prescription database with an acceptable accuracy. We have proven that these communities present the disease fields.

Background: Data mining as an integral part of the knowledge discovery in database (KDD) has gained significant attention over the past few years. By and large, data mining is the process of finding interesting structures in a considerably voluminous amount of data. Owing to its methods and algorithms supporting variable types of data, the data mining approach has been applied in many scientific areas, including the healthcare industry. Regarding this matter, in this paper, we elaborate on the latest papers, including data miningtechniques and algorithms in the healthcare field of research. Results: We present a data mining review based on the newest researches. Afterward, we categorize data mining papers in healthcare based on supervised and unsupervised learning paradigms as well as classifying them in terms of their applications in the healthcare domain. Conclusions: In every healthcare application, we propose some summary points of the papers. At last, we delve into the absence and hence, the necessity of existing some novel methods in healthcare domains in this researches.

No abstract

Background and Aim: The existence of an intellectual structure for every field is essential for managers and scholars. Intellectual structures provide a comprehensive map of knowledge that can guide researchers and managers to have a better view of their fields. Besides, with high-speed and massive amounts of data and information generation, reading and surveying of all resources are severely tricky. Intellectual maps solve this problem and make a situation for control and monitoring this voluminous and high-speed generated data. Epidemiology is regarded as one of the exciting fields which many researchers focused on it. A study of the structure and criteria of different epidemiological fields has not been done yet. Indeed, there is no serious effort for knowledge discovery of hidden information on epidemiological texts. Methods: In this paper, in order to survey this field, an intellectual structure is provided using co-word analysis. Utilizing co-word analysis discloses relationships and structure among research subjects and topics in a field. Results: Finally, four main clusters were determined, namely: genetic (with 30.53% of surveyed papers), illness (29.47%), modeling (23.16%), and prevention (16.84%). Conclusion: According to epidemiology co-word network, epidemiology area has not been studied from enough different areas, especially from novel technologies.

Background and Aim: One of the statistical methods used to analyze the time-to-event medical data is survival analysis. In survival models, the response variable is time to the occurrence of an event. The main characteristic of survival data is the existence of censored data. When we have the distribution of survival time, we can use parametric methods. Among the important  and popular distributions that can  be used, we can mention the Weibull distribution. If the data derives from a heterogeneous population, simple parametric models (such as Weibull) would not fit the data appropriately. One of the methods which have been introduced to overcome this problem is the use of mixture models.   Methods: To assess the validity of the two-component Weibull mixture model, we use a simulation method on heterogeneous survival data. For this purpose, data with different sample sizes were produced in a batch of 1000. Then, the validity of the model is checked using root mean square error (RMSE) criterion   Results: It is obtained that increasing the sample size would decrease the RMSE in the parameters. However the maximum observed RMSE in all the parameters was negligible.   Conclusion: The Bayesian Weibull mixture model was a proper fit for the heterogeneous survival data