List of Articles Neural Network

• Open Access Article
  • Abstract Page
  • Full-Text
  
  1 - Automatic Sepration of Learnrs in Learning Groups Based on Identifying Learning Style from Their Behavior in Learning Environment
  mohammad sadegh rezaei gholamali montazer
  Automatic identification of learners groups based on similarity of learning style improves e-learning systems from the viewpoint of learning adaptation and collaboration among learners. In this paper, a new system is proposed for identifying groups of learners, who have More

  Automatic identification of learners groups based on similarity of learning style improves e-learning systems from the viewpoint of learning adaptation and collaboration among learners. In this paper, a new system is proposed for identifying groups of learners, who have similar learning style, by using learners’ behavior information in an e-learning environment. Proposed clustering method for separation of learners is developed based on ART neural network structure and Snap-Drift neural network learning process. This artificial network enables us to identify learners groups in uncertain group separation parameters, without knowing appropriate number of groups.  The results of an empirical evaluation of the proposed method, which are based on two criteria, “Davies-Bouldin” and “Purity and Gathering”, indicate that our proposed method outperforms other clustering methods in terms of accuracy. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  2 - Modified orthogonal chaotic colonial competition algorithm and its application in improving pattern recognition in multilayer perceptron neural network
  Payman Moallem mehrdad sadeghi hariri MAHDI hashemi
  20.1001.1.27170411.1397.10.35.1.7
  Despite the success of the Colonial Competition Algorithm (ICA) in solving optimization problems, this algorithm still suffers from repeated entrapment in the local minimum and low convergence speed. In this paper, a new version of this algorithm, called Modified Orthog More

  Despite the success of the Colonial Competition Algorithm (ICA) in solving optimization problems, this algorithm still suffers from repeated entrapment in the local minimum and low convergence speed. In this paper, a new version of this algorithm, called Modified Orthogonal Chaotic Colonial Competition (COICA), is proposed. In the policy of absorbing the proposed version, each colony seeks the space to move towards the colonizer through the definition of a new orthogonal vector. Also, the possibility of selecting powerful empires is defined through the boltzmann distribution function, and the selection operation is performed through the roulette wheel method. The proposed multilevel perceptron neural network (MLP) algorithm is used to classify standard datasets, including ionosphere and sonar. To evaluate the performance of this algorithm and to evaluate the generalizability of the trained neural network with the proposed version, the K-Fold cross-validation method has been used. The results obtained from the simulations confirm the reduction of network training error as well as the improved generalizability of the proposed algorithm. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  3 - Comparing A Hybridization of Fuzzy Inference System and Particle Swarm Optimization Algorithm with Deep Learning to Predict Stock Prices
  Majid Abdolrazzagh-Nezhad mahdi kherad
  20.1001.1.27170411.1398.11.41.5.0
  Predicting stock prices by data analysts have created a great business opportunity for a wide range of investors in the stock markets. But the fact is difficulte, because there are many affective economic factors in the stock markets that they are too dynamic and compl More

  Predicting stock prices by data analysts have created a great business opportunity for a wide range of investors in the stock markets. But the fact is difficulte, because there are many affective economic factors in the stock markets that they are too dynamic and complex. In this paper, two models are designed and implemented to identify the complex relationship between 10 economic factors on the stock prices of companies operating in the Tehran stock market. First, a Mamdani Fuzzy Inference System (MFIS) is designed that the fuzzy rules set of its inference engine is found by the Particle Swarm Optimization Algorithm (PSO). Then a Deep Learning model consisting of 26 neurons is designed wiht 5 hidden layers. The designed models are implemented to predict the stock prices of nine companies operating on the Tehran Stock Exchange. The experimental results show that the designed deep learning model can obtain better results than the hybridization of MFIS-PSO, the neural network and SVM, although the interperative ability of the obtained patterns, more consistent behavior with much less variance, as well as higher convergence speed than other models can be mentioned as significant competitive advantages of the MFIS-PSO model Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  4 - Converting protein sequence to image for classification with convolutional neural network
  reza ahsan mansour ebrahimi dianat dianat
  20.1001.1.27170414.1399.12.43.11.2
  Since methods for sequencing machine learning sequences were not successful in classifying healthy and cancerous proteins, it is imperative to find a way to represent these sequences to classify healthy and ill individuals with deep learning approaches. In this study di More

  Since methods for sequencing machine learning sequences were not successful in classifying healthy and cancerous proteins, it is imperative to find a way to represent these sequences to classify healthy and ill individuals with deep learning approaches. In this study different methods of protein sequence representation for classification of protein sequence of healthy individuals and leukemia have been studied. Results showed that conversion of amino acid letters to one-dimensional feature vectors in classification of 2 classes was not successful and only one disease class was detected. By changing the feature vector to colored numbers, the accuracy of the healthy class recognition was slightly improved. The binary protein sequence representation method was more efficient than the previous methods with the initiative of sequencing the sequences in both one-dimensional and two-dimensional (image by Gabor filtering). Protein sequence representation as binary image was classified by applying Gabor filter with 100% accuracy of the protein sequence of healthy individuals and 98.6% protein sequence of those with leukemia. The findings of this study showed that the representation of protein sequence as binary image by applying Gabor filter can be used as a new effective method for representation of protein sequences for classification Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  5 - A Novel Method based on the Cocomo model to increase the accuracy of software projects effort estimates
  mahdieh salari vahid khatibi amid khatibi
  20.1001.1.27170411.1399.12.45.3.3
  It is regarded as a crucial task in a software project to estimate the criteria, and effort estimation in the primary stages of software development is thus one of the most important challenges involved in management of software projects. Incorrect estimation can lead t More

  It is regarded as a crucial task in a software project to estimate the criteria, and effort estimation in the primary stages of software development is thus one of the most important challenges involved in management of software projects. Incorrect estimation can lead the project to failure. It is therefore a major task in efficient development of software projects to estimate software costs accurately. Therefore, two methods were presented in this research for effort estimation in software projects, where attempts were made to provide a way to increase accuracy through analysis of stimuli and application of metaheuristic algorithms in combination with neural networks. The first method examined the effect of the cuckoo search algorithm in optimization of the estimation coefficients in the COCOMO model, and the second method was presented as a combination of neural networks and the cuckoo search optimization algorithm to increase the accuracy of effort estimation in software development. The results obtained on two real-world datasets demonstrated the proper efficiency of the proposed methods as compared to that of similar methods. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  6 - An Intelligent Model for Multidimensional Personality Recognition of Users using Deep Learning Methods
  Hossein Sadr fatemeh mohades deilami morteza tarkhan
  20.1001.1.27170414.1400.13.47.6.2
  Due to the significant growth of textual information and data generated by humans on social networks, there is a need for systems that can automatically analyze the data and extract valuable information from them. One of the most important textual data is people's opini More

  Due to the significant growth of textual information and data generated by humans on social networks, there is a need for systems that can automatically analyze the data and extract valuable information from them. One of the most important textual data is people's opinions about a particular topic that are expressed in the form of text. Text published by users on social networks can represent their personality. Although machine learning based methods can be considered as a good choice for analyzing these data, there is also a remarkable need for deep learning based methods to overcome the complexity and dispersion of content and syntax of textual data during the training process. In this regard, the purpose of this paper is to employ deep learning based methods for personality recognition. Accordingly, the convolutional neural network is combined with the Adaboost algorithm to consider the possibility of using the contribution of various filter lengths and gasp their potential in the final classification via combining various classifiers with respective filter sizes using AdaBoost. The proposed model was conducted on Essays and YouTube datasets. Based on the empirical results, the proposed model presented superior performance compared to other existing models on both datasets. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  7 - Sentiment analysis for stock market predection with deep neural network: A case study for international corporate stock database
  hakimeh mansour Saeedeh Momtazi Kamran Layeghi
  20.1001.1.27170414.1400.13.47.7.3
  Emotional analysis is used as one of the main pillars in various fields such as financial management, marketing and economic changes forecasting in different countries. In order to build an emotion analyzer based on users' opinions on social media, after extracting impo More

  Emotional analysis is used as one of the main pillars in various fields such as financial management, marketing and economic changes forecasting in different countries. In order to build an emotion analyzer based on users' opinions on social media, after extracting important features between words by convolutional layers, we use LSTM layers to establish the relationship behind the sequence of words and extract the important features of the text. With discovery of new features extracted by LSTM, the ability of the proposed model to classify the stock values of companies increases. This article is based on the data of Nguyen et al. (2015) and uses only the emotional information of people in social networks to predict stocks. Given that we categorize each user's message into one of the emotional classes "Strong Buy", "Buy", "Hold", "Sell", "Strong Sell", this model can predict the stock value of the next day, whether it will be high or low. The proposed structure consisted of 21 layers of neural networks consisting of convolutional neural networks and long short-term memory network. These networks were implemented to predict the stock markets of 18 companies. Although some of the previously presented models have used for emotion analysis to predict the capital markets, the advanced hybrid methods have not been performed in deep networks with a good forecasting accuracy. The results were compared with 8 baseline methods and indicate that the performance of the proposed method is significantly better than other baselines. For daily forecasts of stocks changes, it resulted in 19.80% improvement in the prediction accuracy, compared with the deep CNN, and 24.50% and 23.94% improvement compared with the models developed by Nguyen et al. (2015) and Derakhshan et al. (2019), respectively. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  8 - Using Sentiment Analysis and Combining Classifiers for Spam Detection in Twitter
  mehdi salkhordeh haghighi Aminolah Kermani
  20.1001.1.27170411.1401.14.51.10.6
  The welcoming of social networks, especially Twitter, has posed a new challenge to researchers, and it is nothing but spam. Numerous different approaches to deal with spam are presented. In this study, we attempt to enhance the accuracy of spam detection by applying one More

  The welcoming of social networks, especially Twitter, has posed a new challenge to researchers, and it is nothing but spam. Numerous different approaches to deal with spam are presented. In this study, we attempt to enhance the accuracy of spam detection by applying one of the latest spam detection techniques and its combination with sentiment analysis. Using the word embedding technique, we give the tweet text as input to a convolutional neural network (CNN) architecture, and the output will detect spam text or normal text. Simultaneously, by extracting the suitable features in the Twitter network and applying machine learning methods to them, we separately calculate the Tweeter spam detection. Eventually, we enter the output of both approaches into a Meta Classifier so that its output specifies the final spam detection or the normality of the tweet text. In this study, we employ both balanced and unbalanced datasets to examine the impact of the proposed model on two types of data. The results indicate an increase in the accuracy of the proposed method in both datasets. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  9 - Multi-level ternary quantization for improving sparsity and computation in embedded deep neural networks
  Hosna Manavi Mofrad Seyed Ali ansarmohammadi Mostafa Salehi
  Deep neural networks (DNNs) have achieved great interest due to their success in various applications. However, the computation complexity and memory size are considered to be the main obstacles for implementing such models on embedded devices with limited memory and co More

  Deep neural networks (DNNs) have achieved great interest due to their success in various applications. However, the computation complexity and memory size are considered to be the main obstacles for implementing such models on embedded devices with limited memory and computational resources. Network compression techniques can overcome these challenges. Quantization and pruning methods are the most important compression techniques among them. One of the famous quantization methods in DNNs is the multi-level binary quantization, which not only exploits simple bit-wise logical operations, but also reduces the accuracy gap between binary neural networks and full precision DNNs. Since, multi-level binary can’t represent the zero value, this quantization does’nt take advantage of sparsity. On the other hand, it has been shown that DNNs are sparse, and by pruning the parameters of the DNNs, the amount of data storage in memory is reduced while computation speedup is also achieved. In this paper, we propose a pruning and quantization-aware training method for multi-level ternary quantization that takes advantage of both multi-level quantization and data sparsity. In addition to increasing the accuracy of the network compared to the binary multi-level networks, it gives the network the ability to be sparse. To save memory size and computation complexity, we increase the sparsity in the quantized network by pruning until the accuracy loss is negligible. The results show that the potential speedup of computation for our model at the bit and word-level sparsity can be increased by 15x and 45x compared to the basic multi-level binary networks. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  10 - Stock market prediction using optimized grasshopper optimization algorithm and time series algorithms
  Vahid Safari dehnavi masoud shafiee
  20.1001.1.27170411.1401.14.53.8.8
  Stock market prediction serves as an attractive and challenging field for researchers in financial markets. Many of the models used in stock market prediction are not able to predict accurately or these models require a large amount of input data, which increases the vo More

  Stock market prediction serves as an attractive and challenging field for researchers in financial markets. Many of the models used in stock market prediction are not able to predict accurately or these models require a large amount of input data, which increases the volume of networks and learning complexity, all of which ultimately reduce the accuracy of forecasting. This article proposes a method for forecasting the stock market that can effectively predict the stock market. In this paper, the past market price is used to reduce the volume of input data and this data is placed in a regressor model. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  11 - Presenting the ICT Policies Implementation Model of the 6th Development Using the Neural Network Method
  Nazila Mohammadi Gholamreza   Memarzadeh Tehran Sedigheh Tootian Isfahani
  It is inevitable to properly manage the implementation of information and communication technology policies in a planned way in order to improve the country's position in the fields of science and technology. The purpose of this research is to provide a model of the eff More

  It is inevitable to properly manage the implementation of information and communication technology policies in a planned way in order to improve the country's position in the fields of science and technology. The purpose of this research is to provide a model of the effective factors on the implementation of Iran's ICT policies with the help of the neural network technique and based on Giddens' constructive theory. From the point of view of conducting it, this research is of a survey type and based on the purpose, it is of an applied type because it is trying to use the results of the research in the Ministry of Communication and Information Technology and the Iranian Telecommunications Company. Data collection is based on library and field method. The tool for collecting information is research researcher-made questionnaire. The statistical population of the research is information and communication technology experts at the headquarters of Iran Telecommunication Company (810 people), of which 260 people were randomly selected as a sample based on Cochran's formula. MATLAB software was used for data analysis. According to the findings, the best combination for development is when all input variables are considered at the same time, and the worst case is when the infrastructure development variable is ignored, and the most important based on network sensitivity analysis is related to infrastructure development and the least important is related to content supply. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  12 - Predicting the workload of virtual machines in order to reduce energy consumption in cloud data centers using the combination of deep learning models
  Zeinab Khodaverdian Hossein Sadr Mojdeh Nazari Soleimandarabi Seyed Ahmad Edalatpanah
  20.1001.1.27170411.1402.15.55.9.0
  Cloud computing service models are growing rapidly, and inefficient use of resources in cloud data centers leads to high energy consumption and increased costs. Plans of resource allocation aiming to reduce energy consumption in cloud data centers has been conducted usi More

  Cloud computing service models are growing rapidly, and inefficient use of resources in cloud data centers leads to high energy consumption and increased costs. Plans of resource allocation aiming to reduce energy consumption in cloud data centers has been conducted using live migration of Virtual Machines (VMs) and their consolidation into the small number of Physical Machines (PMs). However, the selection of the appropriate VM for migration is an important challenge. To solve this issue, VMs can be classified according to the pattern of user requests into Delay-sensitive (Interactive) or Delay-Insensitive classes, and thereafter suitable VMs can be selected for migration. This is possible by virtual machine workload prediction .In fact, workload predicting and predicting analysis is a pre-migration process of a virtual machine. In this paper, In order to classification of VMs in the Microsoft Azure cloud service, a hybrid model based on Convolution Neural Network (CNN) and Gated Recurrent Unit (GRU) is proposed. Microsoft Azure Dataset is a labeled dataset and the workload of virtual machines in this dataset are in two labeled Delay-sensitive (Interactive) or Delay-Insensitive. But the distribution of samples in this dataset is unbalanced. In fact, many samples are in the Delay-Insensitive class. Therefore, Random Over-Sampling (ROS) method is used in this paper to overcome this challenge. Based on the empirical results, the proposed model obtained an accuracy of 94.42 which clearly demonstrates the superiority of our proposed model compared to other existing models. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  13 - Design and implementation of a survival model for patients with melanoma based on data mining algorithms
  farinaz sanaei Seyed Abdollah  Amin Mousavi Abbas Toloie Eshlaghy ali rajabzadeh ghotri
  10.61186/jict.44093.15.57.1
  20.1001.1.27170411.1402.15.57.1.6
  Background/Purpose: Among the most commonly diagnosed cancers, melanoma is the second leading cause of cancer-related death. A growing number of people are becoming victims of melanoma. Melanoma is also the most malignant and rare form of skin cancer. Advanced cases of More

  Background/Purpose: Among the most commonly diagnosed cancers, melanoma is the second leading cause of cancer-related death. A growing number of people are becoming victims of melanoma. Melanoma is also the most malignant and rare form of skin cancer. Advanced cases of the disease may cause death due to the spread of the disease to internal organs. The National Cancer Institute reported that approximately 99,780 people were diagnosed with melanoma in 2022, and approximately 7,650 died. Therefore, this study aims to develop an optimization algorithm for predicting melanoma patients' survival. Methodology: This applied research was a descriptive-analytical and retrospective study. The study population included patients with melanoma cancer identified from the National Cancer Research Center at Shahid Beheshti University between 2008 and 2013, with a follow-up period of five years. An optimization model was selected for melanoma survival prognosis based on the evaluation metrics of data mining algorithms. Findings: A neural network algorithm, a Naïve Bayes network, a Bayesian network, a combination of decision tree and Naïve Bayes network, logistic regression, J48, and ID3 were selected as the models used in the national database. Statistically, the studied neural network outperformed other selected algorithms in all evaluation metrics. Conclusion: The results of the present study showed that the neural network with a value of 0.97 has optimal performance in terms of reliability. Therefore, the predictive model of melanoma survival showed a better performance both in terms of discrimination power and reliability. Therefore, this algorithm was proposed as a melanoma survival prediction model. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  14 - An Intrusion Detection System based on Deep Learning for CAN Bus
  Fatemeh Asghariyan Mohsen Raji
  20.1001.1.27170411.1402.15.57.2.7
  In recent years, with the advancement of automotive electronics and the development of modern vehicles with the help of embedded systems and portable equipment, in-vehicle networks such as the controller area network (CAN) have faced new security risks. Since the CAN bu More

  In recent years, with the advancement of automotive electronics and the development of modern vehicles with the help of embedded systems and portable equipment, in-vehicle networks such as the controller area network (CAN) have faced new security risks. Since the CAN bus lacks security systems such as authentication and encryption to deal with cyber-attacks, the need for an intrusion detection system to detect attacks on the CAN bus seem to be very necessary. In this paper, a deep adversarial neural network (DACNN) is proposed to detect various types of security intrusions in CAN buses. For this purpose, the DACNN method, which is an extension of the CNN method using adversarial learning, detects intrusion in three stages; In the first stage, CNN acts as a feature descriptor and the main features are extracted, and in the second stage, the discriminating classifier classifies these features and finally, the intrusion is detected using the adversarial learning. In order to show the efficiency of the proposed method, a real open source dataset was used in which the CAN network traffic on a real vehicle during message injection attacks is recorded on a real vehicle. The obtained results show that the proposed method performs better than other machine learning methods in terms of false negative rate and error rate, which is less than 0.1% for DoS and drive gear forgery attack and RPM forgery attack while this rate is less than 0.5% for fuzzy attack. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  15 - Improvement of intrusion detection system on Industrial Internet of Things based on deep learning using metaheuristic algorithms
  mohammadreza zeraatkarmoghaddam majid ghayori
  20.1001.1.27170411.1402.15.57.12.7
  Due to the increasing use of industrial Internet of Things (IIoT) systems, one of the most widely used security mechanisms is intrusion detection system (IDS) in the IIoT. In these systems, deep learning techniques are increasingly used to detect attacks, anomalies or i More

  Due to the increasing use of industrial Internet of Things (IIoT) systems, one of the most widely used security mechanisms is intrusion detection system (IDS) in the IIoT. In these systems, deep learning techniques are increasingly used to detect attacks, anomalies or intrusions. In deep learning, the most important challenge for training neural networks is determining the hyperparameters in these networks. To overcome this challenge, we have presented a hybrid approach to automate hyperparameter tuning in deep learning architecture by eliminating the human factor. In this article, an IDS in IIoT based on convolutional neural networks (CNN) and recurrent neural network based on short-term memory (LSTM) using metaheuristic algorithms of particle swarm optimization (PSO) and Whale (WOA) is used. This system uses a hybrid method based on neural networks and metaheuristic algorithms to improve neural network performance and increase detection rate and reduce neural network training time. In our method, considering the PSO-WOA algorithm, the hyperparameters of the neural network are determined automatically without the intervention of human agent. In this paper, UNSW-NB15 dataset is used for training and testing. In this research, the PSO-WOA algorithm has use optimized the hyperparameters of the neural network by limiting the search space, and the CNN-LSTM neural network has been trained with this the determined hyperparameters. The results of the implementation indicate that in addition to automating the determination of hyperparameters of the neural network, the detection rate of are method improve 98.5, which is a good improvement compared to other methods. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  16 - Multi-Level Ternary Quantization for Improving Sparsity and Computation in Embedded Deep Neural Networks
  Hosna Manavi Mofrad ali ansarmohammadi Mostafa Salehi
  Deep neural networks (DNNs) have achieved great interest due to their success in various applications. However, the computation complexity and memory size are considered to be the main obstacles for implementing such models on embedded devices with limited memory and co More

  Deep neural networks (DNNs) have achieved great interest due to their success in various applications. However, the computation complexity and memory size are considered to be the main obstacles for implementing such models on embedded devices with limited memory and computational resources. Network compression techniques can overcome these challenges. Quantization and pruning methods are the most important compression techniques among them. One of the famous quantization methods in DNNs is the multi-level binary quantization, which not only exploits simple bit-wise logical operations, but also reduces the accuracy gap between binary neural networks and full precision DNNs. Since, multi-level binary can’t represent the zero value, this quantization does not take advantage of sparsity. On the other hand, it has been shown that DNNs are sparse, and by pruning the parameters of the DNNs, the amount of data storage in memory is reduced while computation speedup is also achieved. In this paper, we propose a pruning and quantization-aware training method for multi-level ternary quantization that takes advantage of both multi-level quantization and data sparsity. In addition to increasing the accuracy of the network compared to the binary multi-level networks, it gives the network the ability to be sparse. To save memory size and computation complexity, we increase the sparsity in the quantized network by pruning until the accuracy loss is negligible. The results show that the potential speedup of computation for our model at the bit and word-level sparsity can be increased by 15x and 45x compared to the basic multi-level binary networks. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  17 - Presenting the ICT Policies Implementation Model of the 6th Development Using the Neural Network Method
  Nazila Mohammadi Gholamreza  Memarzadeh sedigheh tootian
  It is inevitable to properly manage the implementation of information and communication technology policies in a planned way in order to improve the country's position in the fields of science and technology. The purpose of this research is to provide a model of the eff More

  It is inevitable to properly manage the implementation of information and communication technology policies in a planned way in order to improve the country's position in the fields of science and technology. The purpose of this research is to provide a model of the effective factors on the implementation of Iran's ICT policies by the neural network technique and based on Giddens' constructive theory. From the point of view of conducting it, this research is of a survey type and based on the purpose, it is of an applied type because it is trying to use the results of the research in the Ministry of Communication and Information Technology and the Iranian Telecommunications Company. Data collection is based on library and field method. The tool for collecting information is researcher-made questionnaire. The statistical population of the research is ICT experts at the headquarters of Iran Telecommunication Company (810 people), of which 260 people were randomly selected as a sample based on Cochran's formula. MATLAB software was used for data analysis. According to the findings, the best combination for development is when all input variables are considered at the same time, and the worst case is when the infrastructure development variable is ignored, and the most important based on network sensitivity analysis is related to infrastructure development and the least important is related to content supply. Manuscript profile