Neural Networks are models of biological neural structure, so the scientist, engineers & mathematicians etc. try to make an intellectual abstraction with the help of neural network which would enable a computer work in a similar fashion in which the human brain works. Here we use a specific type of neural network called “Holographic Neural Network” (HNN), for stock price prediction. HNN takes in the input through Stimulus Vector and gives output through Response Vector. Each element in HNN is associated with a confidence & magnitude value, for this the input given should be in polar form of complex numbers. The results predicted by HNN are compared to results predicted by Regression method.
{"title":"Application of Holographic Neural Network for Stock Price Prediction","authors":"Vaishnavi R. Kunkoliker","doi":"10.1109/ICMLC.2010.42","DOIUrl":"https://doi.org/10.1109/ICMLC.2010.42","url":null,"abstract":"Neural Networks are models of biological neural structure, so the scientist, engineers & mathematicians etc. try to make an intellectual abstraction with the help of neural network which would enable a computer work in a similar fashion in which the human brain works. Here we use a specific type of neural network called “Holographic Neural Network” (HNN), for stock price prediction. HNN takes in the input through Stimulus Vector and gives output through Response Vector. Each element in HNN is associated with a confidence & magnitude value, for this the input given should be in polar form of complex numbers. The results predicted by HNN are compared to results predicted by Regression method.","PeriodicalId":423912,"journal":{"name":"2010 Second International Conference on Machine Learning and Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128658146","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}
As one of the popular evolutionary algorithms, differential evolution (DE) shows outstanding convergence rate on continuous optimization problems. But prematurity probably still occurs in classical DE when using relatively small population, which is discussed in this paper. Considering that large population may significantly raise the computational effort, we propose a modified DE using smaller population (DESP) by introducing extra disturbance to its mutation operation. In addition, an adaptive adjustment scheme is designed to control the disturbance intensity according to the improvement during the evolution. To test the performance of DESP, two groups of experiments are conducted. The results show that DESP outperforms DE in terms of convergence rate and accuracy.
{"title":"Differential Evolution Using Smaller Population","authors":"Xuan Ren, Zhi-zhao Chen, Zhen Ma","doi":"10.1109/ICMLC.2010.9","DOIUrl":"https://doi.org/10.1109/ICMLC.2010.9","url":null,"abstract":"As one of the popular evolutionary algorithms, differential evolution (DE) shows outstanding convergence rate on continuous optimization problems. But prematurity probably still occurs in classical DE when using relatively small population, which is discussed in this paper. Considering that large population may significantly raise the computational effort, we propose a modified DE using smaller population (DESP) by introducing extra disturbance to its mutation operation. In addition, an adaptive adjustment scheme is designed to control the disturbance intensity according to the improvement during the evolution. To test the performance of DESP, two groups of experiments are conducted. The results show that DESP outperforms DE in terms of convergence rate and accuracy.","PeriodicalId":423912,"journal":{"name":"2010 Second International Conference on Machine Learning and Computing","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133946087","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}
Video compression plays an important role in video signal processing, transmission and storage. Since the available bandwidth for transmission is very limited, Multimedia Applications such as video conferencing, video on demand, video telephony and remote sensing are not possible without compression. A lot of video compression techniques have been developed and the video signal transmission has followed at data rates below 64kbps. Wavelet transform based motion compensated video codec performs better compression in order to meet the rate and distortion constraint in video transmission for the available bandwidth than the block based techniques, which are followed in standard video transmissions such as H.261 and H.263. But the efficiency of those technique’s depends on the way in which it estimates and compensates the object motions in the video sequence. Wavelet based embedded image coder is quite attractive in modern multimedia applications. Wavelet transform, bit plane coding and other techniques make embedded image coder practical and also provide efficient compression. In this paper, we have proposed a novel video coding using swarm intelligence in dual tree complex wavelet transform for video coding. The 3-D DDWT is an attractive video representation because it isolates motion along different directions in separate subbands. However, it is an over-complete transform with redundancy, which is going to be eliminated by choosing optimal subbands with the help of PSO. The proposed video codec does not require motion compensation and provides better performance than the 3D SPIHT (Embedded type)codec, both objectively and subjectively, and the coder allows full scalability in spatial, temporal and quality dimensions.
{"title":"Video Coding Technique Using Swarm Intelligence in 3-D Dual Tree Complex Wavelet Transform","authors":"M. Thamarai, R. Shanmugalakshmi","doi":"10.1109/ICMLC.2010.39","DOIUrl":"https://doi.org/10.1109/ICMLC.2010.39","url":null,"abstract":"Video compression plays an important role in video signal processing, transmission and storage. Since the available bandwidth for transmission is very limited, Multimedia Applications such as video conferencing, video on demand, video telephony and remote sensing are not possible without compression. A lot of video compression techniques have been developed and the video signal transmission has followed at data rates below 64kbps. Wavelet transform based motion compensated video codec performs better compression in order to meet the rate and distortion constraint in video transmission for the available bandwidth than the block based techniques, which are followed in standard video transmissions such as H.261 and H.263. But the efficiency of those technique’s depends on the way in which it estimates and compensates the object motions in the video sequence. Wavelet based embedded image coder is quite attractive in modern multimedia applications. Wavelet transform, bit plane coding and other techniques make embedded image coder practical and also provide efficient compression. In this paper, we have proposed a novel video coding using swarm intelligence in dual tree complex wavelet transform for video coding. The 3-D DDWT is an attractive video representation because it isolates motion along different directions in separate subbands. However, it is an over-complete transform with redundancy, which is going to be eliminated by choosing optimal subbands with the help of PSO. The proposed video codec does not require motion compensation and provides better performance than the 3D SPIHT (Embedded type)codec, both objectively and subjectively, and the coder allows full scalability in spatial, temporal and quality dimensions.","PeriodicalId":423912,"journal":{"name":"2010 Second International Conference on Machine Learning and Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116279425","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}
Najla Krichen Masmoudi, C. Rekik, M. Djemel, N. Derbel
This paper presents a method to compute optimal control strategies of discrete large scale nonlinear systems by using hierarchical fuzzy systems. The method is based on the decomposition principle of the global system into interconnected subsystems becoming easier to study. Then, the differential dynamic programming procedure is applied in order to obtain the rule basis. After that, we construct limpid-hierarchical Mamdani fuzzy system in order to compute optimal control laws, for each subsystem. Simulation results of a rotary crane show that the proposed method yields to satisfactory performances. The robustness of the proposed approach is verified.
{"title":"Optimal Control for Discrete Large Scale Nonlinear Systems Using Hierarchical Fuzzy Systems","authors":"Najla Krichen Masmoudi, C. Rekik, M. Djemel, N. Derbel","doi":"10.1109/ICMLC.2010.32","DOIUrl":"https://doi.org/10.1109/ICMLC.2010.32","url":null,"abstract":"This paper presents a method to compute optimal control strategies of discrete large scale nonlinear systems by using hierarchical fuzzy systems. The method is based on the decomposition principle of the global system into interconnected subsystems becoming easier to study. Then, the differential dynamic programming procedure is applied in order to obtain the rule basis. After that, we construct limpid-hierarchical Mamdani fuzzy system in order to compute optimal control laws, for each subsystem. Simulation results of a rotary crane show that the proposed method yields to satisfactory performances. The robustness of the proposed approach is verified.","PeriodicalId":423912,"journal":{"name":"2010 Second International Conference on Machine Learning and Computing","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121237547","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}
M. Rahiman, Aswathy Shajan, A. Elizabeth, M. Divya, G. M. Kumar, M. Rajasree
Recently Indian Handwritten character recognition is getting much more attention and researchers are contributing a lot in this field. But Malayalam, a South Indian language has very less works in this area and needs further attention. This paper focuses on an efficient algorithm for recognizing the handwritten Malayalam characters. Malayalam OCR is a complex task owing to the various character scripts available and more importantly the difference in ways in which the characters are written. The dimensions are never the same and may be never mapped on to a square grid unlike English characters. Here we propose an algorithm which can accept the scanned image of handwritten characters as input and to produce the editable Malayalam characters in a predefined format as output without applying any resizing or skeletonization methods but still can produce much accurate results. Characters are grouped in to different classes based on their HLH intensity patterns. These patterns are separated from the image and fed for recognition. Algorithm is tested for 4 sets of samples ranging 661 letters in the noiseless environment and produces an accuracy of 88%.
{"title":"Isolated Handwritten Malayalam Character Recognition Using HLH Intensity Patterns","authors":"M. Rahiman, Aswathy Shajan, A. Elizabeth, M. Divya, G. M. Kumar, M. Rajasree","doi":"10.1109/ICMLC.2010.8","DOIUrl":"https://doi.org/10.1109/ICMLC.2010.8","url":null,"abstract":"Recently Indian Handwritten character recognition is getting much more attention and researchers are contributing a lot in this field. But Malayalam, a South Indian language has very less works in this area and needs further attention. This paper focuses on an efficient algorithm for recognizing the handwritten Malayalam characters. Malayalam OCR is a complex task owing to the various character scripts available and more importantly the difference in ways in which the characters are written. The dimensions are never the same and may be never mapped on to a square grid unlike English characters. Here we propose an algorithm which can accept the scanned image of handwritten characters as input and to produce the editable Malayalam characters in a predefined format as output without applying any resizing or skeletonization methods but still can produce much accurate results. Characters are grouped in to different classes based on their HLH intensity patterns. These patterns are separated from the image and fed for recognition. Algorithm is tested for 4 sets of samples ranging 661 letters in the noiseless environment and produces an accuracy of 88%.","PeriodicalId":423912,"journal":{"name":"2010 Second International Conference on Machine Learning and Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131662447","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}
Medicine has always benefited from the technology. Artificial Neural Networks is currently the promising area of interest to solve medical problems. Diagnosis of diabetes is one of the most challenging problems in machine learning. This medical data set is seldom complete. Artificial neural networks require complete set of data for an accurate classification. The system explains how the pre-processing procedure and missing values influence the data set during the classification. The implemented system compares various missing value techniques and pre-processing techniques. Some combinations prove the real influence of these techniques. A classifier has applied to Pima Indian Diabetes dataset and the results were improved tremendously when using certain combination of preprocessing and missing value techniques. The experimental system achieves an excellent classification accuracy of 99% which is best than before.
{"title":"Impact of Preprocessing for Diagnosis of Diabetes Mellitus Using Artificial Neural Networks","authors":"T. Jayalskshmi, A. Santhakumaran","doi":"10.1109/ICMLC.2010.65","DOIUrl":"https://doi.org/10.1109/ICMLC.2010.65","url":null,"abstract":"Medicine has always benefited from the technology. Artificial Neural Networks is currently the promising area of interest to solve medical problems. Diagnosis of diabetes is one of the most challenging problems in machine learning. This medical data set is seldom complete. Artificial neural networks require complete set of data for an accurate classification. The system explains how the pre-processing procedure and missing values influence the data set during the classification. The implemented system compares various missing value techniques and pre-processing techniques. Some combinations prove the real influence of these techniques. A classifier has applied to Pima Indian Diabetes dataset and the results were improved tremendously when using certain combination of preprocessing and missing value techniques. The experimental system achieves an excellent classification accuracy of 99% which is best than before.","PeriodicalId":423912,"journal":{"name":"2010 Second International Conference on Machine Learning and Computing","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132375587","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}
In this paper the effect of training neural network BFSK demodulator with noisy data (sent by transmitter and affected by channel) is discussed and the results is compared with predefined noiseless data bits. Distributed time-delay neural network is selected and get trained by both noisy and noiseless data bits. Simulations show that training a neural network demodulator by predetermined data bits sent by transmitter (noisy data) helps demodulator detect data bits with less error. That is because noisy data can give the neural network demodulator some information about channel behavior and environmental noise and consequently it can help receiver to detect data bits intelligently. Matlab simulations in an AWGN channel prove the idea.
{"title":"Improving ANN BFSK Demodulator Performance with Training Data Sequence Sent by Transmitter","authors":"M. Amini, E. Balarastaghi","doi":"10.1109/ICMLC.2010.28","DOIUrl":"https://doi.org/10.1109/ICMLC.2010.28","url":null,"abstract":"In this paper the effect of training neural network BFSK demodulator with noisy data (sent by transmitter and affected by channel) is discussed and the results is compared with predefined noiseless data bits. Distributed time-delay neural network is selected and get trained by both noisy and noiseless data bits. Simulations show that training a neural network demodulator by predetermined data bits sent by transmitter (noisy data) helps demodulator detect data bits with less error. That is because noisy data can give the neural network demodulator some information about channel behavior and environmental noise and consequently it can help receiver to detect data bits intelligently. Matlab simulations in an AWGN channel prove the idea.","PeriodicalId":423912,"journal":{"name":"2010 Second International Conference on Machine Learning and Computing","volume":"221 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121565721","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}
In this paper the analysis of wall static pressure variation has been done with fuzzy logic approach to have smooth flow in the duct. Here there are three area ratio choosen for the enlarged duct, 2.89, 6.00 and 10.00. The primary pressure ratio is taken as 2.65 and cavity aspect ratio is taken as 1 and 2. The study is analysed for length to diameter ratio of 1,2,4 and 6. The nozzles used are De Laval type and with a Mach number of 1.74 and 2.23 and conical nozzles having Mach numbers of 1.58 and 2.06. The analysis based on fuzzy logic theory indicates that the length to diameter ratio of 1 is sufficient for smooth flow development if only the basis of wall static pressure variations is considered.
{"title":"Wall Static Pressure Variation in Sudden Expansion in Cylindrical Ducts with Supersonic Flow: A Fuzzy Logic Approach","authors":"K. Pandey","doi":"10.1109/ICMLC.2010.74","DOIUrl":"https://doi.org/10.1109/ICMLC.2010.74","url":null,"abstract":"In this paper the analysis of wall static pressure variation has been done with fuzzy logic approach to have smooth flow in the duct. Here there are three area ratio choosen for the enlarged duct, 2.89, 6.00 and 10.00. The primary pressure ratio is taken as 2.65 and cavity aspect ratio is taken as 1 and 2. The study is analysed for length to diameter ratio of 1,2,4 and 6. The nozzles used are De Laval type and with a Mach number of 1.74 and 2.23 and conical nozzles having Mach numbers of 1.58 and 2.06. The analysis based on fuzzy logic theory indicates that the length to diameter ratio of 1 is sufficient for smooth flow development if only the basis of wall static pressure variations is considered.","PeriodicalId":423912,"journal":{"name":"2010 Second International Conference on Machine Learning and Computing","volume":"241 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2010-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115220874","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：