Artificial Neural Network

Organizations expose to financial risk that can lead to bankruptcy and loss of business is increased nowadays. This may leads to discontinuity in operations, increased legal fees, administrative costs and other indirect costs. Accordingly, the purpose of this study was to predict the financial crisis of Tehran Stock Exchange using neural network and genetic algorithm. This research is descriptive and practical and in order to collect data Stock Exchange database software has been used. For data analysis, we used artificial neural network in base form and artificial neural network mix with genetic algorithm. In addition for methods comparison, determination coefficient, Mean squared error and Root-mean square error have been used. The result of study shows that the best artificial neural network is a network with a hidden layer and eight neurons in the layer. This network could predict 97.7 percent of healthy and bankrupt companies correctly for test data. Furthermore the best mixed neural network with genetic algorithm is a network with 400 replications and population size 50, one layer and eight neurons which could correctly predict 100% of healthy and bankrupt companies. Finally, comparison of results of two methods shows that the best method for predicting financial crisis is mixed neural network with genetic algorithm.

Cash flow is one of the critical resources in the economic unit and the balance between available cash and cash needs is the most important factor in economic health. Since judgments of many stakeholders such as investors and shareholders about the position of the economic unit are based on liquidity situation, so predicting future cash flow is crucial. In this research, the impact of cash and accrual items on cash flow forecasts has been studied. Providing a proper model to predict operating cash flows and review some important characteristics of cash flow forecasting regression models, using a multilayer perceptron and determining the best model by using accrual regression model variables for predicting cash flows. For this purpose, 287 firms listed in Tehran Stock Exchange during 2008 to 2017 were studied; Linear and nonlinear regression, correlation coefficient and artificial neural network statistical methods have been used for data analysis and predictive power of powers was compared by using the sum of squared prediction error and coefficient of determination. Results showed that the accrual regression model can predict future cash flows better than other tested models and among corporate characteristics, the highest correlation belongs to sales volatility and firm size with accrual regression models. On the other hand, results of fitting different neural network models indicate that two structures with 8 and 11 hidden nodes are the best models to predict cash flows.

Network data envelopment analysis models assess efficiency of decision-making unit and its sections using historical data but fail to measure efficiency of its units and their internal stages in the future. In this paper we aim to measure efficiency of stages of bank branches and obtain efficiency trend of stages during the time, then to estimate their efficiency in the future therefore we can be aware of stages inefficiency before occurrence and prevent them. First, a two-stage structure including deposit collection and loan giving was designed for bank branches using literature review and comments of experts. Human forces and fixed assets were considered as input variables of the first stage, deposit as mediator variable, delayed claims as interim variable, and loan amount as output variable of the second stage. Then, a dynamic network data envelopment analysis model was formulated and stages efficiency were obtained for 16 consecutive periods. Therefore, efficiency trend of stages was obtained during the time. In the following, efficiency of various stages of branches were estimated using artificial neural network and some recommendations are provided according to obtained amounts in order to prevent inefficiency before occurrence.

The liberalization of natural gas markets and the emergence of gas hubs in recent decades have shifted the natural gas trade from the regional to the global trade. The growth and maturity of these hubs have weakened the previously established relationship between the natural gas price and the prices of crude oil and petroleum products. Therefore, predicting the price of gas as a strategic commodity has become more important for different countries. Using the neural network method, this paper attempts to provide a model for the monthly prediction of natural gas price. Based on the time series data from 2012 to April 2019 as neural network input, this model predicts the prices in five hubs and natural gas exchange centers in Europe. Based on the R2 performance evaluation index of 98% in the neural network model fitted based on the aforementioned data series, the neural network model has an acceptable performance in predicting the natural gas price. The results of this study show that using the artificial neural network (ANN) method, the gas prices in the European gas hubs, which are located in European Country, can be predicted with a high accuracy.

With the emergence of the World Wide Web, Electronic Commerce (E-commerce) has been growing rapidly in the past two decades. Intelligent agents play the main role in making the negotiation between different entities automatically.  Automated negotiation allows resolving opponent agents' mutual concerns to reach an agreement without the risk of losing individual profits. However, due to the unknown information about the opponent's strategies, automated negotiation is difficult. The main challenge is how to reveal the optimal information about the opponent's strategy during the negotiation process to propose the best counter-offer. In this paper, we design a buyer agent which can automatically negotiate with the opponent using artificial intelligence techniques and machine learning methods. The proposed buyer agent is designed to learn the opponent's strategies during the negotiation process using four methods: "Bayesian Learning", "Kernel Density Estimation", "Multilayer Perceptron Neural Network", and "Nonlinear Regression". Experimental results show that the use of machine learning methods increases the negotiation efficiency, which is measured and evaluated by parameters such as the rate agreement (RA), average buyer utility (ABU), average seller utility (ASU), average rounds (AR). Rate agreement and average buyer utility have increased from 58% to 74% and 90% to 94%, respectively, and average rounds have decreased from 10% to 0.04%.

Stock market forecasting has attracted so many researchers and investors that many studies have been done in this field. These studies have led to the development of many predictive methods, the most widely used of which are machine learning-based methods. In machine learning-based methods, loss function has a key role in determining the model weights. In this study a new loss function is introduced, that has some special features, making the investing in the stock market more accurate and profitable than other popular techniques. To assess its accuracy, a two-stage experiment has been designed using data of Tehran Stock market. In the first part of the experiment, we select the most accurate algorithm among some of the well-known machine learning algorithms based on artificial neural network, ANN, support vector machine, SVM. In the second stage of the experiment, the various popular loss functions are compared with the proposed one. As a result, we introduce a new neural network using a new loss function, which is trained based on genetic algorithm. This network has been shown to be more accurate than other well-known and common networks such as long short-term memory (LSTM) for both train and test data.

This study was an attempt to evaluate the progress of capital market efficiency in Iran. Optimal resource allocation and micro and macro investments play a key role in the capital market. The capital market's main task is to circulate capital and allocate resources efficiently and optimally. The main task of this market is to flow capital and allocate resources efficiently and optimally. Is there a regular pattern for determining the stock price? Market efficiency gains significance as it is important to know what factor or factors are effective in determining the price of the stock in the stock market or whether there is a regular pattern for determining the price of a stock. Thus, this study examined the efficiency of the capital market in Iran. In this regard, the researchers used the daily data of the total index of the Tehran Stock Exchange for 2008-2017. Artificial neural network and time series training tests were used to perform the test. The test results showed weak efficiency in the Tehran Stock Exchange and this inefficiency did not change significantly compared to the first period. In other words, in the Tehran Stock Market, one can predict returns using artificial intelligence.

The application of machine learning technologies for cancer detection purposes are rising due to their ever-increasing accuracy. Melanoma is one of the most common types of skin cancer. Detection of melanoma in the early stages can significantly prevent illness and fetal death. The application of innovative machine learning technology is highly relevant and valuable due to medical practitioners' difficulty in early-stage diagnoses. This paper provides an open-source tutorial on the performance of an algorithm that helps to diagnose melanoma by extracting features from dermatoscopic images and their classification. First, we used a Dull-Razor preprocessing method to remove extra details such as hair. Next, histogram adjustments and lighting thresholds were used to increase the contrast and select lesion boundaries. After using a threshold, a binary-classified version of image was obtained, and the boundary of the lesion was determined. As a result, the features from skin tissue were extracted. Finally, a comparative study was conducted between three methods which are Artificial Neural Network (ANN), Support Vector Machine (SVM) and K-Nearest Neighbor (KNN). The results show that ANN could achieve better accuracy (83.5%). In order to mitigate the biases in existing studies, the source code of this research is available at hadi-naghavipour.com/ml to serve aspiring researchers for improvement, correction and learning and provide a guideline for technology manager practitioners.

Objective: The purpose of this study was to modeling the role of organizational ethics atmospheres and work-family conflict in cyber loafing. Methods: Methodology of the research in terms of the strategy was quantitative-field and in terms of analytical, was descriptive. The statistical population of this study included all staff of Universities affiliated with the Ministry of Science at four levels of international, national, regional and local levels of performance across the country. The research data were collected from 430 staff members who were selected by multi-stage cluster sampling method. The sample size according to the Kregci-Morgan model and with error α = 0/05, was considered 430 persons. To collect data Cyber Loafing Questionnaire of Bella et al (2006), Organizational Ethical Atmospheres Questionnaire of Victor and Cullen (1988) and work-family Conflict Questionnaire of DuBrin (1985) was used. Validity of the tools was confirmed by the professors of education and psychology. Data were analyzed by artificial neural network with multilayer perceptron (MPL) method. Results: The results showed modeling the organizational factors affecting on cyber loafing has an input layer with seven units and a hidden layer with one unit and the artificial neural network is well able to predict the jumps and the process of cyber loafing of academic staff based on organizational factors of organizational ethics atmospheres and work-family conflict. Conclusion: It is necessary to pay attention to these components and their effects on the organization human resource behavior and avoiding deviant behavior.

No unique method has been so far specified for determining the number of neurons in hidden layers of Multi-Layer Perceptron (MLP) neural networks used for prediction. The present research is intended to optimize the number of neurons using two meta-heuristic procedures namely genetic and hill climbing algorithms. The data used in the present research for prediction are consumption data of water subscribers in Fasa City of Fars Province (Iran) between the years 2010 to 2013. Ultimately, using the respective data set, the data of the subsequent year 2014 can be predicted. In the present research it was observed that the mean square errors of per data (MSEPD) for the abovementioned algorithms are less than 0.2, indicating a high performance in the neural networks’ prediction. Correlation coefficients using genetic and hill climbing algorithms were respectively equal to 0.891 and 0.759. Thus, GA was able to leave a better effect on optimization of neural network.