Search results
1 – 4 of 4No study has investigated the effects of different parameters on publication bias in meta-analyses using a machine learning approach. Therefore, this study aims to evaluate the…
Abstract
Purpose
No study has investigated the effects of different parameters on publication bias in meta-analyses using a machine learning approach. Therefore, this study aims to evaluate the impact of various factors on publication bias in meta-analyses.
Design/methodology/approach
An electronic questionnaire was created according to some factors extracted from the Cochrane Handbook and AMSTAR-2 tool to identify factors affecting publication bias. Twelve experts were consulted to determine their opinion on the importance of each factor. Each component was evaluated based on its content validity ratio (CVR). In total, 616 meta-analyses comprising 1893 outcomes from PubMed that assessed the presence of publication bias in their reported outcomes were randomly selected to extract their data. The multilayer perceptron (MLP) technique was used in IBM SPSS Modeler 18.0 to construct a prediction model. 70, 15 and 15% of the data were used for the model's training, testing and validation partitions.
Findings
There was a publication bias in 968 (51.14%) outcomes. The established model had an accuracy rate of 86.1%, and all pre-selected nine variables were included in the model. The results showed that the number of databases searched was the most important predictive variable (0.26), followed by the number of searches in the grey literature (0.24), search in Medline (0.17) and advanced search with numerous operators (0.13).
Practical implications
The results of this study can help clinical researchers minimize publication bias in their studies, leading to improved evidence-based medicine.
Originality/value
To the best of the author’s knowledge, this is the first study to model publication bias using machine learning.
Details
Keywords
Hossein Motahari-Nezhad, Maryam Shekofteh and Maryam Kazerani
The purpose of this study is to assess the e-readiness status of libraries in the Shahid Beheshti University of Medical Sciences in terms of four dimensions – human resources…
Abstract
Purpose
The purpose of this study is to assess the e-readiness status of libraries in the Shahid Beheshti University of Medical Sciences in terms of four dimensions – human resources, electronical infrastructure, network services and programs and enhancers of the networked world.
Design/methodology/approach
The study population consists of 11 libraries of the Shahid Beheshti University of Medical Sciences, including the central library and 10 faculty libraries. The data collection instrument is a questionnaire prepared by the researchers that has been designed on the basis of the “e-readiness assessment of Iranian academic libraries model”. Depending on the respondents there are three parts to the questionnaire: questionnaire for managers, staff and information and communication technology (ICT) officials. Their reliability and validity have been proved.
Findings
The libraries of Shahid Beheshti University of Medical Sciences had an average to high status in terms of “human resources” with a score of 2.32, “electronic infrastructure” with a score of 2.48, “network services and programs” with a score of 2.09 and “networked world enhancers” with a score of 2.37 out of 4. In total, these libraries had an average to high status in terms of e-readiness, with a score of 2.29.
Originality/value
The findings of this study can help the library administrators and officials of Shahid Beheshti University of Medical Sciences to plan improvements to the situation of ICT.
Details
Keywords
Hossein Motahari-Nezhad, Maryam Shekofteh and Maryam Andalib-Kondori
This study aims to investigate the characteristics, as well as the purpose and posts of the COVID-19 Facebook groups.
Abstract
Purpose
This study aims to investigate the characteristics, as well as the purpose and posts of the COVID-19 Facebook groups.
Design/methodology/approach
A systematic search for COVID-19 Facebook groups was conducted on June 1, 2020. Characteristics of the groups were examined using descriptive statistics. Mann-Whitney test was used to study the differences between groups. The study of the most popular groups’ posts was also carried out using the content analysis method.
Findings
The groups had a combined membership of 2,729,061 users. A total of 147,885 posts were received. There were about approximately 60% public groups. A high percentage of the groups (86.5%) had descriptions. The results showed a significant relationship between the groups’ description status and the number of members (p-value = 0.016). The majority of COVID-19 Facebook groups (56%) were created to meet their members’ information needs. The highest number of studied posts were related to vaccination (35.2%), followed by curfew rules (19.6%) and symptoms (10.6%).
Originality/value
Translating these insights into policies and practices will put policymakers and health-care providers in a stronger position to make better use of Facebook groups to support and enhance public knowledge about COVID-19.
Details
Keywords
Mehrdad Fadaei PellehShahi, Sohrab Kordrostami, Amir Hossein Refahi Sheikhani and Marzieh Faridi Masouleh
Predicting the final status of an ongoing process or a subsequent activity in a process is an important aspect of process management. Semi-structured business processes cannot be…
Abstract
Purpose
Predicting the final status of an ongoing process or a subsequent activity in a process is an important aspect of process management. Semi-structured business processes cannot be predicted by precise and mathematical methods. Therefore, artificial intelligence is one of the successful methods. This study aims to propose a method that is a combination of deep learning methods, in particular, the recurrent neural network and Markov chain.
Design/methodology/approach
The proposed method applies the BestFirst algorithm for the search section and the Cfssubseteval algorithm for the feature comparison section. This study focuses on the prediction systems of social insurance and tries to present a method that is less costly in providing real-world results based on the past history of an event.
Findings
The proposed method is simulated with real data obtained from Iranian Social Security Organization, and the results demonstrate that using the proposed method increases the memory utilization slightly more than the Markov method; however, the CPU usage time has dramatically decreased in comparison with the Markov method and the recurrent neural network and has, therefore, significantly increased the accuracy and efficiency.
Originality/value
This research tries to provide an approach capable of producing the findings closer to the real world with fewer time and processing overheads, given the previous records of an event and the prediction systems of social insurance.
Details