Data mining and related topics
Saman Haratizadeh; Fatemeh Rezaee
Abstract
Purpose: Selection of the best stocks for the portfolio as well as allocating the optimal amount of capital per stock in the portfolio are serious challenges in investing in the stock market. The use of machine learning capacities in the process of optimal capital allocation among portfolio assets has ...
Read More
Purpose: Selection of the best stocks for the portfolio as well as allocating the optimal amount of capital per stock in the portfolio are serious challenges in investing in the stock market. The use of machine learning capacities in the process of optimal capital allocation among portfolio assets has received less attention and usually, the same weight is assigned to portfolio stocks or traditional risk assessment methods are used to divide capital between portfolio stocks. The common disadvantage of these methods is that they all use simple and inflexible mechanisms to estimate the performance of a set. The purpose of this paper is to show for the first time, that machine learning can be used to create a more effective mechanism for estimating performance, which leads to a more efficient allocation of capital to portfolio stocks.Methodology: Our proposed framework, uses two predictive models based on machine learning. In the first step, stocks historical information is used in a return forecasting model, then based on the predicted returns, the appropriate stocks of the portfolio are selected. In the second step, a separate forecasting model predicts portfolio returns by taking into account both the forecasted returns in the first model and the expected risk of the stocks. At the end based on the predicted return of the numerous random portfolios, the appropriate weight for each asset is selected.Findings: Comparing the returns of adjusted portfolios with this model and adjusted portfolios with other portfolio optimization methods shows the superiority of the proposed model.Originality/Value: In this paper, by using machine learning models, the process of selecting the appropriate stock of the portfolio and allocating capital among the candidate stocks is done optimally.
Data mining and related topics
hassan khademi zare; atena moghimi; mohammadsaleh owlia; Davood Shishebori
Abstract
Nowadays, it is obvious that media has a dramatic role in people's lives. Among all types of media, TV can still be powerful if it tries to know its audience and uses creative management. A management that considers the interests of the media and the audience as one is effective.To achieve this goal, ...
Read More
Nowadays, it is obvious that media has a dramatic role in people's lives. Among all types of media, TV can still be powerful if it tries to know its audience and uses creative management. A management that considers the interests of the media and the audience as one is effective.To achieve this goal, we should look for solutions to increase media influence by analyzing the audience and the programs together.In this article, we used innovative joint clustering of audiences based on demographic characteristics and characteristics of television programs. Solutions are provided for members of each cluster in order to increase media influence. The data was obtained from a researcher-made questionnaire and a sample of 390 related experts and people of Yazd.According to the demands of the audience during watching peaks, the managers of Yazd’s local TV channel have to review their broadcast schedule and use the solutions provided in this article on their agenda. Evaluating the quality of clustering shows its suitable structure. The proposed solutions have been validated according to the opinions of media experts and the degree of result’s compliance with the sources related to the research topic.In this article, due to a new technique called joint clustering, the audience is clustered hierarchically and simultaneously based on demographic characteristics and television programs. in addition, the solutions are provided for members in each cluster to increase media influence.
Data mining and related topics
Ramin Safa; Peyman Bayat; Leila Moghtader
Abstract
Purpose: While diagnosing mental disorders in traditional approaches relies on questionnaires, interviews, and clinical trials, automated screening tools can take a shorter path. These tools can be developed as innovative evaluation techniques, decision support systems, and prevention strategies to help ...
Read More
Purpose: While diagnosing mental disorders in traditional approaches relies on questionnaires, interviews, and clinical trials, automated screening tools can take a shorter path. These tools can be developed as innovative evaluation techniques, decision support systems, and prevention strategies to help susceptible individuals. Due to the tendency of people to share thoughts and feelings on social platforms, microblogging data contains valuable information that can be analyzed to identify users’ mental states. This study describes a roadmap for data analysis in the field in question.Methodology: In the first part of this paper, concepts such as electronic mental health and microblogging platforms are introduced. And their conceptual relationship is discussed by providing explanations about data science and social data analysis. Next, the prediction of disorder in social platforms is described separately. Finally, by reviewing related works and open issues, we explain how data collection, pre-processing, and analysis are done using different features of real-world data.Findings: By experimental analysis, this study shows that the extracted features from the users’ profiles have a significant effect on predicting mental health symptoms, and even with the information extracted from the users’ public profiles, the mental state can be predicted with reasonable accuracy.Originality/Value: In this study, automatic analysis of social data to investigate the psychological signals is described, and in the implementation section, it is founded that the symptoms can be followed in almost all the studied features.
Data mining and related topics
Hossein Sadr; Mojdeh Nazari Soleimandarabi; Zeinab Khodaverdian
Abstract
Purpose: Automatic short answer grading is known as the task of automatic assessment of answers based on natural language using computation methods and machine learning algorithms. The proliferation of large-scale intelligent education systems and the importance of assessment as a key factor in the education ...
Read More
Purpose: Automatic short answer grading is known as the task of automatic assessment of answers based on natural language using computation methods and machine learning algorithms. The proliferation of large-scale intelligent education systems and the importance of assessment as a key factor in the education process have increased the need for highly flexible automated systems for scoring exams.Methodology: While in the process of automatic short answer grading, student's answer is compared to an ideal response and scoring is done based on their similarity, semantic relatedness and similarity measures can also be employed for this aim. To this end, several semantic relatedness and similarity measures are firstly compared in application of short answer grading. In the following, a method for improving the performance of short answer grading systems based on semantic relatedness and similarity measures which leverages students' answers with the highest score as feedback is proposed.Findings: In order to evaluate the performance of semantic and similarity relatedness methods in application of automatic short answer grading and the prposed model, various experiments were concucted on Mohler and Mihalcea dataset that contains 7 questions and 630 answers.Originality/Value: Based on the empirical experiments not only semantic relatedness and similarity measures have great efficiency in automatic short answer grading but also using students' answers as feedback can considerably improve the accuracy and performance of semantic relatedness and similarity measures for this task.
Data mining and related topics
Fatemeh Mirsaeedi; hamidreza koosha; Mohammad Ghodoosi
Abstract
Survey academic performance by educational data mining is one of the most important issues in the field of educational management and researchers focus on it. The purpose of this study is to present an experimental method for appropriate algorithm selection in predicting students' academic status in ...
Read More
Survey academic performance by educational data mining is one of the most important issues in the field of educational management and researchers focus on it. The purpose of this study is to present an experimental method for appropriate algorithm selection in predicting students' academic status in two and three classes. Two-class database predicts the admission or rejection of students in the course, while the database of the three classes, in addition to admission or rejection, identifies students who are prone and elite. Using the previous articles in the field of educational data mining and experts' opinions, factors that effect on academic performance of students were identified and database was compiled based on them. After optimization of parameters and implementation of different algorithms, the performance scores of the algorithms were calculated using paired t-test based on three indexes include of accuracy, f-measure, and ROC, algorithms were compared by TOPSIS and VIKOR methods. In the two-class mode, Support Vector Machine algorithm in TOPSIS with value of 0.999115 and VIKOR with value of zero has shown the best performance. In the multi-class mode, the Logistic Regression algorithm in TOPSIS and VIKOR in turns with values 0.9986044 and 0.0009798 performances better than other algorithms. The proposed method can be used as a tool for selecting algorithm that has the best pergormance in educational data mining. Because choosing the algorithm to achieve accurate and exact results is very effective and can be taken into account in the process of counseling and preventing students' academic failure