{"title":"Regularized Compression of A Noisy Blurred Image","authors":"P. Favati, G. Lotti, O. Menchi, F. Romani","doi":"10.5121/IJCSA.2016.6601","DOIUrl":"https://doi.org/10.5121/IJCSA.2016.6601","url":null,"abstract":"","PeriodicalId":39465,"journal":{"name":"International Journal of Computer Science and Applications","volume":"23 1","pages":"1-13"},"PeriodicalIF":0.0,"publicationDate":"2016-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81104381","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Problems from the inside of an organization’s perimeters are a significant threat, since it is very difficult to differentiate them from outside activity. In this dissertation, evaluate an insider threat detection motto on its ability to detect different type of scenarios that have not previously been identify or contemplated by the developers of the system. We show the ability to detect a large variety of insider threat scenario instances We report results of an ensemble-based, unsupervised technique for detecting potential insider threat, insider threat scenarios that robustly achieves results. We explore factors that contribute to the success of the ensemble method, such as the number and variety of unsupervised detectors and the use of existing knowledge encoded in scenario based detectors made for different known activity patterns. We report results over the entire period of the ensemble approach and of ablation experiments that remove the scenario-based detectors.
{"title":"Detecting Unknown Insider Threat Scenarios","authors":"M. S. Lodhi, Rahul Kaul","doi":"10.5121/IJCSA.2016.6602","DOIUrl":"https://doi.org/10.5121/IJCSA.2016.6602","url":null,"abstract":"Problems from the inside of an organization’s perimeters are a significant threat, since it is very difficult to differentiate them from outside activity. In this dissertation, evaluate an insider threat detection motto on its ability to detect different type of scenarios that have not previously been identify or contemplated by the developers of the system. We show the ability to detect a large variety of insider threat scenario instances We report results of an ensemble-based, unsupervised technique for detecting potential insider threat, insider threat scenarios that robustly achieves results. We explore factors that contribute to the success of the ensemble method, such as the number and variety of unsupervised detectors and the use of existing knowledge encoded in scenario based detectors made for different known activity patterns. We report results over the entire period of the ensemble approach and of ablation experiments that remove the scenario-based detectors.","PeriodicalId":39465,"journal":{"name":"International Journal of Computer Science and Applications","volume":"10 1","pages":"15-21"},"PeriodicalIF":0.0,"publicationDate":"2016-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75244782","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The most emerging prominent sensor network applications collect data from sensor nodes and monitors periodically. Resource constraint Sensor motes sense the environment and transit data to the remote sink via multiple hops. Minimum energy dissipation and secure data transmission are crucial to such applications. This paper delivers an energy efficient, lifetime improving, secure periodic Data Gathering scheme that is a hybrid of heuristic path establishment and secure data transmission. This protocol uses artificial intelligence (AI) based A* heuristic search algorithm to establish energy efficient admissible optimal path to sink in terms of high residual energy, minimum hop counts and high link quality. This scheme also adopts block encryption Rivest Cipher (RC6) Algorithm to secure the transmission of packets. This code and speed optimized block encryption provides confidentiality against critical data and consumes less energy for encryption. This proposed method increases the network lifetime there by reducing the total traffic load. Evaluation of performance analysis of this algorithm using Network Simulator (NS2) shows the superiority of the proposed scheme.
{"title":"ENERGY EFFICIENT, LIFETIME IMPROVING AND SECURE PERIODIC DATA COLLECTION PROTOCOL FOR WIRELESS SENSOR NETWORKS","authors":"P. Anuja, Dr. Raju Shanmugam","doi":"10.5121/IJCSA.2016.6301","DOIUrl":"https://doi.org/10.5121/IJCSA.2016.6301","url":null,"abstract":"The most emerging prominent sensor network applications collect data from sensor nodes and monitors periodically. Resource constraint Sensor motes sense the environment and transit data to the remote sink via multiple hops. Minimum energy dissipation and secure data transmission are crucial to such applications. This paper delivers an energy efficient, lifetime improving, secure periodic Data Gathering scheme that is a hybrid of heuristic path establishment and secure data transmission. This protocol uses artificial intelligence (AI) based A* heuristic search algorithm to establish energy efficient admissible optimal path to sink in terms of high residual energy, minimum hop counts and high link quality. This scheme also adopts block encryption Rivest Cipher (RC6) Algorithm to secure the transmission of packets. This code and speed optimized block encryption provides confidentiality against critical data and consumes less energy for encryption. This proposed method increases the network lifetime there by reducing the total traffic load. Evaluation of performance analysis of this algorithm using Network Simulator (NS2) shows the superiority of the proposed scheme.","PeriodicalId":39465,"journal":{"name":"International Journal of Computer Science and Applications","volume":"26 1","pages":"1-11"},"PeriodicalIF":0.0,"publicationDate":"2016-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86643334","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Virtual communities are becoming increasingly popular, particularly on the internet, as a means for like-minded individuals to meet, share or gain access to the information they are most interested in quickly and efficiently, from other individuals. The concept of a community of interest can be supported in a virtual community in order to bring the appropriate parties together, to share their knowledge with each other. The objective of our work is to investigate a community abstraction approach to agent based knowledge management.
{"title":"A Novel Clustering Algorithm For Coverage A Large Scale In Wireless Sensor Networks","authors":"Wassim Jerbi, Abderrahmen Guermazi, H. Trabelsi","doi":"10.5121/IJCSA.2016.6201","DOIUrl":"https://doi.org/10.5121/IJCSA.2016.6201","url":null,"abstract":"Virtual communities are becoming increasingly popular, particularly on the internet, as a means for like-minded individuals to meet, share or gain access to the information they are most interested in quickly and efficiently, from other individuals. The concept of a community of interest can be supported in a virtual community in order to bring the appropriate parties together, to share their knowledge with each other. The objective of our work is to investigate a community abstraction approach to agent based knowledge management.","PeriodicalId":39465,"journal":{"name":"International Journal of Computer Science and Applications","volume":"5 1","pages":"01-18"},"PeriodicalIF":0.0,"publicationDate":"2016-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76748856","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-01-01DOI: 10.12783/IJCSA.2016.0501.01
A. Sundaram, G. Ramesh
Solar power is a promising solution in recent environmental aspects so several methods of generation were implemented in past to reach advances in generation capacity and critical conditions of solar. This paper deals with Non-inverting Four Switch buck-boost converter (NFSBBC) which is presented to reach advances in generation with respect to solar condition. The advances of generation include maximum power point tracking using incremental conductance, reliable buck boost operation and voltage stability ratio. Incremental conductance can track rapidly on solar power than perturb and observer method, so proposed power generation draws a continuous power across DC-Link inverter drive for wide speed operation of brushless DC motor. A new phase current infusion is implemented to estimate flux, torque and control of speed in wide range without external infusion and sensing loops in present sensor less vector control. 110W/50V capacity of solar is implemented for present solar power generation; output capacity of converter reaches from 25-100V using buck boost operation. 25V/1500 rpm capacity of brushless DC Motor is controlled using proposed sensor less vector using current infusing logic.
{"title":"An Advance Solar Power Generation and Control of Brushless DC Motor Using Phase Current Infusion of Sensor Less Vector Control","authors":"A. Sundaram, G. Ramesh","doi":"10.12783/IJCSA.2016.0501.01","DOIUrl":"https://doi.org/10.12783/IJCSA.2016.0501.01","url":null,"abstract":"Solar power is a promising solution in recent environmental aspects so several methods of generation were implemented in past to reach advances in generation capacity and critical conditions of solar. This paper deals with Non-inverting Four Switch buck-boost converter (NFSBBC) which is presented to reach advances in generation with respect to solar condition. The advances of generation include maximum power point tracking using incremental conductance, reliable buck boost operation and voltage stability ratio. Incremental conductance can track rapidly on solar power than perturb and observer method, so proposed power generation draws a continuous power across DC-Link inverter drive for wide speed operation of brushless DC motor. A new phase current infusion is implemented to estimate flux, torque and control of speed in wide range without external infusion and sensing loops in present sensor less vector control. 110W/50V capacity of solar is implemented for present solar power generation; output capacity of converter reaches from 25-100V using buck boost operation. 25V/1500 rpm capacity of brushless DC Motor is controlled using proposed sensor less vector using current infusing logic.","PeriodicalId":39465,"journal":{"name":"International Journal of Computer Science and Applications","volume":"17 1","pages":"1-11"},"PeriodicalIF":0.0,"publicationDate":"2016-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81381061","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-01-01DOI: 10.12783/IJCSA.2016.0501.02
Ben Beklisi Kwame Ayawli, G. A. Gershon
The purpose of this study was to design an e-voting system to improve verification and auditing of votes and election results to aid build confidence in the use of e-voting. A top-down design approach was adopted in the design and development of the evoting system. BK Web-based e-voting system after design was implemented in ten different elections in 2012, 2013 and 2014. It was very helpful in ensuring voter’s confidence in e-voting since they were able to verify their votes through Short Message Service, reports generated from the system and the provision of ‘strong room’ facility. The verification procedures demonstrated in this study when applied in e-voting systems would help build voter’s confidence in e-voting systems.
{"title":"BK Web-Based E-Voting System: Ensuring Confidence with Votes and Results Verification","authors":"Ben Beklisi Kwame Ayawli, G. A. Gershon","doi":"10.12783/IJCSA.2016.0501.02","DOIUrl":"https://doi.org/10.12783/IJCSA.2016.0501.02","url":null,"abstract":"The purpose of this study was to design an e-voting system to improve verification and auditing of votes and election results to aid build confidence in the use of e-voting. A top-down design approach was adopted in the design and development of the evoting system. BK Web-based e-voting system after design was implemented in ten different elections in 2012, 2013 and 2014. It was very helpful in ensuring voter’s confidence in e-voting since they were able to verify their votes through Short Message Service, reports generated from the system and the provision of ‘strong room’ facility. The verification procedures demonstrated in this study when applied in e-voting systems would help build voter’s confidence in e-voting systems.","PeriodicalId":39465,"journal":{"name":"International Journal of Computer Science and Applications","volume":"20 1","pages":"12-22"},"PeriodicalIF":0.0,"publicationDate":"2016-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82528670","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-01-01DOI: 10.12783/IJCSA.2016.0502.02
Mesnan Silalahi
This paper describes the ontology construction using a metadesign approach. The approach will constitutes key concepts such as: supporting human-problem interaction, underdesigning for emergent behavior, enabling legitimate peripheral participation, sharing control, promoting mutual learning and support, rewarding and recognizing contributions, and fostering reflective communities. The approach supports for flexibility and extensibility in the process and the used tools. The development of semantic web applications was part of the tools support to follow the metadesign framework. The central actor along the phases are the domain experts who also were respondencts in the survey. Based on the resultant ontology and the survey of the participatory domain-experts in a case-study, this approach proved to effectively build the ontology.
{"title":"A Metadesign Approach in Ontology Construction","authors":"Mesnan Silalahi","doi":"10.12783/IJCSA.2016.0502.02","DOIUrl":"https://doi.org/10.12783/IJCSA.2016.0502.02","url":null,"abstract":"This paper describes the ontology construction using a metadesign approach. The approach will constitutes key concepts such as: supporting human-problem interaction, underdesigning for emergent behavior, enabling legitimate peripheral participation, sharing control, promoting mutual learning and support, rewarding and recognizing contributions, and fostering reflective communities. The approach supports for flexibility and extensibility in the process and the used tools. The development of semantic web applications was part of the tools support to follow the metadesign framework. The central actor along the phases are the domain experts who also were respondencts in the survey. Based on the resultant ontology and the survey of the participatory domain-experts in a case-study, this approach proved to effectively build the ontology.","PeriodicalId":39465,"journal":{"name":"International Journal of Computer Science and Applications","volume":"205 1","pages":"35-43"},"PeriodicalIF":0.0,"publicationDate":"2016-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73766715","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-01-01DOI: 10.12783/IJCSA.2016.0502.01
Tuba Firdaus, Meenakshi Yadav
A number of research have taken place in the field of Wireless Sensor Networks (WSN) as there is continuous need of advancement in the field of wireless communication, digital technology and micro-electro-mechanical systems(MEMS) . So the need of growth of low cost, low power, multifunctional sensor nodes have been required. A Wireless Sensor Network is a collection of sensor nodes that have the capability of sensing any environmental phenomenon, processing that information and then sending that data to the base station. A single sensor node is not capable of capturing desired information from a particular region so a collection of nodes are arranged to get accurate and sufficient result. This collection of sensor nodes along with a base station will collaboratively form a network that is known as Wireless Sensor Network. As limited energy is one of the most important constraint of WSN so it must be assured that it is utilized in most efficient way. Clustering is best approach to remove redundant data transmission to base station. Each cluster has a cluster head that is responsible for transmitting data to base station for that cluster members. Cluster head (CH) collect the data from all members of its cluster and perform aggregation on these data to remove redundancy then send it to base station. So appropriate CH election is very important for improving efficiency. In this thesis we have presented a clustering approach that has taken a heterogeneous environment and uses fuzzy logic to elect CHs more efficiently. We have combined two parameters Distance and Residual Energy and apply fuzzy rules on that to find the priority of a node for being a CH. Simulation shows that using fuzzy logic in SEP (Stability Election Protocol) will improve the energy efficiency by providing better load distribution and utilizing the benefits of heterogeneity of network. We have shown our analysis on two parameters- Number of dead nodes and average energy of nodes
{"title":"Modified Clustering Algorithm for Energy Efficiency Utilizing Fuzzy Logic in WSN (MCF)","authors":"Tuba Firdaus, Meenakshi Yadav","doi":"10.12783/IJCSA.2016.0502.01","DOIUrl":"https://doi.org/10.12783/IJCSA.2016.0502.01","url":null,"abstract":"A number of research have taken place in the field of Wireless Sensor Networks (WSN) as there is continuous need of advancement in the field of wireless communication, digital technology and micro-electro-mechanical systems(MEMS) . So the need of growth of low cost, low power, multifunctional sensor nodes have been required. A Wireless Sensor Network is a collection of sensor nodes that have the capability of sensing any environmental phenomenon, processing that information and then sending that data to the base station. A single sensor node is not capable of capturing desired information from a particular region so a collection of nodes are arranged to get accurate and sufficient result. This collection of sensor nodes along with a base station will collaboratively form a network that is known as Wireless Sensor Network. As limited energy is one of the most important constraint of WSN so it must be assured that it is utilized in most efficient way. Clustering is best approach to remove redundant data transmission to base station. Each cluster has a cluster head that is responsible for transmitting data to base station for that cluster members. Cluster head (CH) collect the data from all members of its cluster and perform aggregation on these data to remove redundancy then send it to base station. So appropriate CH election is very important for improving efficiency. In this thesis we have presented a clustering approach that has taken a heterogeneous environment and uses fuzzy logic to elect CHs more efficiently. We have combined two parameters Distance and Residual Energy and apply fuzzy rules on that to find the priority of a node for being a CH. Simulation shows that using fuzzy logic in SEP (Stability Election Protocol) will improve the energy efficiency by providing better load distribution and utilizing the benefits of heterogeneity of network. We have shown our analysis on two parameters- Number of dead nodes and average energy of nodes","PeriodicalId":39465,"journal":{"name":"International Journal of Computer Science and Applications","volume":"5 1","pages":"24-34"},"PeriodicalIF":0.0,"publicationDate":"2016-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88843401","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-10-01DOI: 10.12783/ijcsa.2015.0402.02
Luciano T. E. Pansanato, A. Rivolli, Douglas Felipe Pereira
The user interaction logging with a Rich Internet Applications can provide valuable information for usability evaluation and user behavior study. However, current research has focused only on issues of implementation and validation of tools for capturing user interaction instead tool usage by web developers. This work presents an evaluation through user testing of a tool called WAUTT for capturing user interaction with Rich Internet Applications. The main goal of this evaluation is to investigate the use of a tool for capturing user interaction with web developers. The results indicate that the participants have good understanding of the tool, found easiness to operate the tool for capturing and analyzing the user interaction, consider its future re-use, and they felt satisfied with the experience. The results also assisted for identifying the user interaction data considered more useful to web developers.
{"title":"An Evaluation with Web Developers of Capturing User Interaction with Rich Internet Applications for Usability Evaluation","authors":"Luciano T. E. Pansanato, A. Rivolli, Douglas Felipe Pereira","doi":"10.12783/ijcsa.2015.0402.02","DOIUrl":"https://doi.org/10.12783/ijcsa.2015.0402.02","url":null,"abstract":"The user interaction logging with a Rich Internet Applications can provide valuable information for usability evaluation and user behavior study. However, current research has focused only on issues of implementation and validation of tools for capturing user interaction instead tool usage by web developers. This work presents an evaluation through user testing of a tool called WAUTT for capturing user interaction with Rich Internet Applications. The main goal of this evaluation is to investigate the use of a tool for capturing user interaction with web developers. The results indicate that the participants have good understanding of the tool, found easiness to operate the tool for capturing and analyzing the user interaction, consider its future re-use, and they felt satisfied with the experience. The results also assisted for identifying the user interaction data considered more useful to web developers.","PeriodicalId":39465,"journal":{"name":"International Journal of Computer Science and Applications","volume":"50 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81618675","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-04-01DOI: 10.12783/ijcsa.2015.0401.02
V. Moyo, K. Sibanda
Artificial Neural Networks (ANNs) have attracted increasing attention from researchers in many fields. They have proved to be one of the most powerful tools in the domain of forecasting and analysis of various time series. The ability to model almost any kind of function regardless of its degree of nonlinearity, positions ANNs as good candidates for predicting and modelling self-similar time series such as TCP/IP traffic. Inspite of this, one of the most difficult and least understood tasks in the design of ANN models is the selection of the most appropriate size of the learning rate. Although some guidance in the form of heuristics is available for the choice of this parameter, none have been universally accepted. In this paper we empirically investigate various sizes of learning rates with the aim of determining the optimum learning rate size for generalization ability of an ANN trained on forecasting TCP/IP network traffic trends. MATLAB Version 7.4.0.287’s Neural Network toolbox version 5.0.2 (R2007a) was used for our experiments. The results are found to be promising in terms of ease of design and use of ANNs. We found from the experiments that, depending on the difficulty of the problem at hand, it is advisable to set the learning rate to 0.1 for the standard Backpropagation algorithm and to either 0.1 or 0.2 if used in conjunction with the momentum term of 0.5 or 0.6. We advise minimal use of the momentum term as it greatly interferes with the training process of ANNs. Although the information obtained from the tests carried out in this paper is specific to the problem considered, it provides users of Back-propagation networks with a valuable guide on the behaviour of ANNs under a wide range of operating conditions. It is important to note that the guidelines accrued from this paper are of an assistive and not necessarily restrictive nature to potential ANN modellers.
人工神经网络(Artificial Neural Networks, ann)越来越受到许多领域研究者的关注。它们已被证明是预测和分析各种时间序列领域中最强大的工具之一。无论其非线性程度如何，几乎可以对任何类型的函数进行建模的能力，使人工神经网络成为预测和建模自相似时间序列(如TCP/IP流量)的良好候选者。尽管如此，人工神经网络模型设计中最困难和最不容易理解的任务之一是选择最合适的学习率大小。虽然对于这个参数的选择有一些启发式的指导，但是没有一个是被普遍接受的。在本文中，我们实证研究了各种学习率的大小，目的是确定用于预测TCP/IP网络流量趋势的人工神经网络泛化能力的最佳学习率大小。我们的实验使用MATLAB Version 7.4.0.287的神经网络工具箱Version 5.0.2 (R2007a)。结果表明，人工神经网络在设计和使用方面是有希望的。我们从实验中发现，根据手头问题的难度，建议将标准反向传播算法的学习率设置为0.1，如果与动量项0.5或0.6结合使用，则设置为0.1或0.2。我们建议尽量少使用动量项，因为它极大地干扰了人工神经网络的训练过程。虽然从本文中进行的测试中获得的信息是特定于所考虑的问题的，但它为反向传播网络的用户提供了关于人工神经网络在广泛工作条件下的行为的有价值的指导。重要的是要注意，从本文中积累的指导方针对潜在的人工神经网络建模者来说是辅助的，而不一定是限制性的。
{"title":"The Generalization Ability of Artificial Neural Networks in Forecasting TCP/IP Traffic Trends: How Much Does the Size of Learning Rate Matter?","authors":"V. Moyo, K. Sibanda","doi":"10.12783/ijcsa.2015.0401.02","DOIUrl":"https://doi.org/10.12783/ijcsa.2015.0401.02","url":null,"abstract":"Artificial Neural Networks (ANNs) have attracted increasing attention from researchers in many fields. They have proved to be one of the most powerful tools in the domain of forecasting and analysis of various time series. The ability to model almost any kind of function regardless of its degree of nonlinearity, positions ANNs as good candidates for predicting and modelling self-similar time series such as TCP/IP traffic. Inspite of this, one of the most difficult and least understood tasks in the design of ANN models is the selection of the most appropriate size of the learning rate. Although some guidance in the form of heuristics is available for the choice of this parameter, none have been universally accepted. In this paper we empirically investigate various sizes of learning rates with the aim of determining the optimum learning rate size for generalization ability of an ANN trained on forecasting TCP/IP network traffic trends. MATLAB Version 7.4.0.287’s Neural Network toolbox version 5.0.2 (R2007a) was used for our experiments. The results are found to be promising in terms of ease of design and use of ANNs. We found from the experiments that, depending on the difficulty of the problem at hand, it is advisable to set the learning rate to 0.1 for the standard Backpropagation algorithm and to either 0.1 or 0.2 if used in conjunction with the momentum term of 0.5 or 0.6. We advise minimal use of the momentum term as it greatly interferes with the training process of ANNs. Although the information obtained from the tests carried out in this paper is specific to the problem considered, it provides users of Back-propagation networks with a valuable guide on the behaviour of ANNs under a wide range of operating conditions. It is important to note that the guidelines accrued from this paper are of an assistive and not necessarily restrictive nature to potential ANN modellers.","PeriodicalId":39465,"journal":{"name":"International Journal of Computer Science and Applications","volume":"35 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2015-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75239244","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：