Constrained optimization problems have a wide range of applications in science, economics, and engineering. In this paper, a neural network model is proposed to solve a class of nonsmooth constrained optimization problems with a nonsmooth convex objective function subje More
Constrained optimization problems have a wide range of applications in science, economics, and engineering. In this paper, a neural network model is proposed to solve a class of nonsmooth constrained optimization problems with a nonsmooth convex objective function subject to nonlinear inequality and affine equality constraints. It is a one-layer non-penalty recurrent neural network based on the differential inclusion. Unlike most of the existing neural network models, there is neither a penalty parameter nor a penalty function in its structure. It has less complexity which leads to the easier implementation of the model for solving optimization problems. The equivalence of optimal solutions set of the main optimization problem and the equilibrium points set of the model is proven. Moreover, the global convergence and the stability of the introduced neural network are shown. Some examples including the L1-norm minimization problem are given and solved by the proposed model to illustrate its performance and effectiveness.
Manuscript profile
In this paper we present a recurrent neural network model to recognize efficient Decision Making Units(DMUs) in Data Envelopment Analysis(DEA). The proposed neural network model is derived from an unconstrained minimization problem. In theoretical aspect, it is shown th More
In this paper we present a recurrent neural network model to recognize efficient Decision Making Units(DMUs) in Data Envelopment Analysis(DEA). The proposed neural network model is derived from an unconstrained minimization problem. In theoretical aspect, it is shown that the proposed neural network is stable in the sense of lyapunov and globally convergent. The proposed model has a single-layer structure. Simulation shows that the proposed model is effective to identify efficient DMUs in DEA.
Manuscript profile
Background and Objective: Proper estimation of the amount of sediment flowing in rivers is important as a data base for many river engineering designs and processes. Qaleh Rudkhan River is one of the most important water basins in the west of Gilan province. The most im More
Background and Objective: Proper estimation of the amount of sediment flowing in rivers is important as a data base for many river engineering designs and processes. Qaleh Rudkhan River is one of the most important water basins in the west of Gilan province. The most important branches of the basin are two branches named Gasht Rudkhan and Ghaleh Rudkhan. The river (Qaleh Rudkhan) is made up of two branches (Heydaralat) and (Nazaralat). Therefore, the purpose of this study was to model the prediction of sediment rate in Qaleh Rudkhan River using long short-term memory neural network (LSTM).
Material and Methodology: In this research, the recorded Debi-sediment statistics related to the statistical period of 1381 to 1395 has been used. These statistics include daily instantaneous Debi in cubic meter per second and daily instantaneous sediment in ton per day, which are measured simultaneously. The data used to model the artificial neural network are Debi-sediment values the accuracy of the predictions was evaluated with three error criteria.
Findings: The three criteria considered are AFE, FFE and n-AFE, respectively.
Discussion and Conclusion: Among these criteria, the FFE criterion showed that the correlation between the model output and the measured sediment data is appropriate. As a result, the LSTM model has the appropriate accuracy to predict the amount of sediment in the two rivers of Qala-e-Rudkhan.
Manuscript profile
Active noise control is based on the destructive interference between the primary noise and generated noise from the secondary source. An antinoise of equal amplitude and opposite phase is generated and combined with the primary noise. In this paper, performance of the More
Active noise control is based on the destructive interference between the primary noise and generated noise from the secondary source. An antinoise of equal amplitude and opposite phase is generated and combined with the primary noise. In this paper, performance of the neural networks is evaluated in active cancellation of sound noise. For this reason, feedforward and recurrent neural networks are designed and trained. After training, performance of the feedforwrad and recurrent networks in noise attenuation are compared. We use Elman network as a recurrent neural network. For simulations, noise signals from a SPIB database are used. In order to compare the networks appropriately, equal number of layers and neurons are considered for the networks. Moreover, training and test samples are similar. Simulation results show that feedforward and recurrent neural networks present good performance in noise cancellation. As it is seen, the ability of recurrent neural network in noise attenuation is better than feedforward network.
Manuscript profile
در این مقاله ما یک شبکه عصبی برای حل مدل CCR در تحلیل پوششی داده ها (DEA) معرفی می کنیم. مدل شبکه عصبی پیشنهادی از یک مسئله مینیمم سازی نامقید حاصل می شود. از دیدگاه تئوری، نشان داده می شود که شبکه عصبی پیشنهادی به مفهوم لیاپانف پایدار و همگرای عمومی به جواب بهینه مدل C More
در این مقاله ما یک شبکه عصبی برای حل مدل CCR در تحلیل پوششی داده ها (DEA) معرفی می کنیم. مدل شبکه عصبی پیشنهادی از یک مسئله مینیمم سازی نامقید حاصل می شود. از دیدگاه تئوری، نشان داده می شود که شبکه عصبی پیشنهادی به مفهوم لیاپانف پایدار و همگرای عمومی به جواب بهینه مدل CCR می باشد. مدل پیشنهادی ساختار تک لایه دارد. با یک مثال عددی موثر بودن مدل پیشنهادی برای حل مدل CCR در DEA نشان داده می شود.
Manuscript profile
Predicting time series of financial markets is a challenging issue in the field of specialized studies of time series and has attracted the attention of many researchers. Due to the presence of big data, this issue has led to the growth of developments in the field of m More
Predicting time series of financial markets is a challenging issue in the field of specialized studies of time series and has attracted the attention of many researchers. Due to the presence of big data, this issue has led to the growth of developments in the field of machine learning models. Due to the importance of this issue, in this study, by using the comparison of different machine learning models such as random forest approaches, support vector machine, artificial neural network and deep learning-based recurrent neural networks to investigate the ability of different machine learning models in prediction. The total index of Tehran Stock Exchange during the period 2013 to 2020 has been discussed. The prediction results of 1, 3 and 6 day courses for the out-of-sample period show that the machine learning method based on the long short-term memory (LSTM) network, a recurrent neural networks, has a better result compared to other models.
Manuscript profile
Forecasting stock prices plays an important role in setting a trading strategy or determining the appropriate timing for buying or selling a stock. Deep Learning (DL) is a type of Artificial Neural Network (ANN) that consists of multiple processing layers and enables hi More
Forecasting stock prices plays an important role in setting a trading strategy or determining the appropriate timing for buying or selling a stock. Deep Learning (DL) is a type of Artificial Neural Network (ANN) that consists of multiple processing layers and enables high-level abstraction to model data. The key advantage of DL models is extracting the good features of input data automatically using a general-purpose learning procedure which is suitable for dynamic time series such as stock price.In this research the ability of Long Short-Term Memory (LSTM) to predict the stock price is studied; moreover, the factors that have significant effects on the stock price is classified and legal and natural person trading is introduced as an important factor which has influence on the stock price. Price data, technical indexes and legal and natural person trading is used as an input data for running the model. The results obtained from LSTM with Dropout layer are better and more stable than simple form of LSTM and RNN models.
Manuscript profile
Today, artificial intelligence has made a big change in the recognition of chart patterns in technical analysis. Although, the emergence of new and complex analytical methods in technical analysis has provided a new challenge for artificial intelligence methods. One of More
Today, artificial intelligence has made a big change in the recognition of chart patterns in technical analysis. Although, the emergence of new and complex analytical methods in technical analysis has provided a new challenge for artificial intelligence methods. One of the popular and complex technical analysis methods is Elliott Wave Theory. On the other hand, the speed of progress of artificial intelligence methods is such that a more powerful method is introduced every time. One of the new and powerful artificial intelligence methods is the deep learning method. Therefore, in this research, a model has been presented to predict the trend of the stock market through the detection of fractal patterns based on Elliott wave theory using deep learning method. In this research, 15 Elliott wave patterns were considered, and then 1002 samples of stock price charts of companies listed on Tehran Stock Exchange were collected and labeled for patterns, and finally entered as input into deep learning algorithm using recurrent neural network model for recognition. In this research, RapidMiner 9.9 software was used and accuracy criteria were used to determine the power of the model. Based on the results, the accuracy of developed model in recognizing patterns is 61%.
Manuscript profile