iosr-JCE   Volume-21 ~ Issue-1 ~ Jan-Feb 2019

Abstract: Artificial intelligence has made its presence felt ubiquitously in different avenues of research and technology wherein the data is large and complex. In the proposed work, to forecast solar irradiation energy; whose structure uses the back-propagation concept and uses the Levenberg Marquardt algorithm is used. The system used hitherto a single layer of hidden neurons. The averaging approach is also been used with 2, 12- and 24-hour averaging scheme so as to increase the accuracy of prediction. The system attains a MAPE of 2.7%. Hence the accuracy attained is 97%. The mean square error has been chosen as the performance function for the proposed algorithm.

Keywords: Solar Energy Prediction, Artificial Neural Network (ANN), Back Propagation, Levenberg-Marquardt (LM) Algorithm, Mean Absolute Percentage Error (MAPE).

[1]. L.Saad Saoud, F.Rahmoune, V.Tourtchine, K.Baddari in the paper "Fully Complex Valued Wavelet Neural Network for Forecasting the Global Solar Irradiation", Springer 2016
[2]. Ministry of New and Renewable energy, Government of India,"Annual Report 2015-16", http://mnre.gov.in, 2016.
[3]. Vishal Sharma, Dazhi Yang, Wilfred Walsh, Thomas Reindl in the paper "Short Term Solar Irradiance Forecasting Using A Mixed Wavelet Neural Network" Elsevier 2016
[4]. Ozgur Kisi, Erdal Uncuoghlu, "Comparison of three backpropagation training algorithms for two case studies," Indian Journal of Engineering & Materials Sciences, Volume 12, pp. 434-442, 2005.
[5]. E.M. Johansson, F.U. Dowla, and D.M. Goodman, "Backpropagation Learning for Multilayer Feed-Forward Neural Networks using The Conjugate Gradient Method," International Journal of Neural Systems, Volume 02, pp. 291-302, 1991..

Abstract: Error correction code (ECC) and built-in selfrepair(BISR) techniques by using redundancies have beenwidely used for improving the yield and reliability of embeddedmemories. The target faults of these two schemes are soft errorsand permanent (hard) faults, respectively. In recent works, thereare also some techniques integrating ECC and BISR to deal withsoft errors and hard defects simultaneously. However, this willcompromise reliability, since some of the ECC protection capabilityis used for repairing single hard faults. To cure this dilemma,we propose an ECC-enhanced BISR (EBISR) technique, whichuses ECC to repair single permanent faults first and sparesfor the remaining faults in the production/power-ON test andrepair stage. However, techniques are proposed to maintain.........

Index Terms—Built-in self-repair (BISR), error correctioncode (ECC), hard repair, reliability, yield.

[1]. Semiconductor Industry Association, "International technology roadmap for semiconductors (ITRS), 2003 edition," Hsinchu, Taiwan, Dec.2003.
[2]. C. Stapper, A. Mclaren, and M. Dreckman, "Yield model for Productivity Optimization of VLSI Memory Chips with redundancy and Partially good Product," IBM Journal of Research and Development, Vol. 24, No. 3, pp. 398-409, May 1980.
[3]. W. K. Huang, Y. H. shen, and F. lombrardi, "New approaches for repairs of memories with redundancy by row/column deletion for yield enhancement," IEEE Transactions on Computer-Aided Design, vol. 9, No. 3, pp. 323-328, Mar. 1990.
[4]. P. Mazumder and Y. S. Jih, "A new built-in self-repair approach to VLSI memory yield enhancement by using neuraltype circuits," IEEE transactions on Computer Aided Design, vol. 12, No. 1, Jan, 1993.
[5]. H. C. Kim, D. S. Yi, J. Y. Park, and C. H. Cho, "A BISR (built-in self repair) circuit for embedded memory with multiple redundancies," VLSI and CAD 6th International Conference, pp. 602-605, Oct. 1999.

Abstract: Frequent itemset mining is one of the important aspects of association rule mining. The primary algorithm based for frequent itemset mining is mostly based on CPU and they generated a large set of items that are required to be kept in memory all the time while processing. In this dissertation thesis, we designed a parallel Eclat algorithm. The algorithm will run on GPU and perform the task of the frequentitemset mining in parallel. The proposed algorithm also uses the optimized candidate representation and the frequent item sets generated are stored in cache memory and are fetched directly from the cache memory. The proposed algorithm runs in parallel and also uses the optimized candidate representation and thus provides better performance than the classical eclat algorithm. Thus, the proposed algorithm runs much faster than the classical eclat algorithm and has better performance than classical eclat algorithm in terms of memory and time..

Keywords: FIM(Frequent Itemset Mining), Equivalent Class, Candidate sets, support count, Eclat

[1]. K. Jiawei, Han Micheline, Data Mining Concepts and Techniques.
[2]. R. Agrawal, T. Imieliski, and A. Swami, Mining association rules between sets of items in large databases,ACM SIGMOD Rec.,22(2), 1993, 207-216.
[3]. S. Dutt, N. Choudhary, and D. Singh, An Improved Apriori Algorithm based on Matrix Data Structure, Global Journal of Computer Science and Technology: C Software & Data Engineering,14(5), 2014.
[4]. U. Grag, and M. Kaur, ECLAT Algorithm for Frequent Itemsets Generation, International Journal of Computer Systems,1(3), 2014,1-4.
[5]. NVIDIA Corporation, "Datasheet: NVIDIA Kepler Next-Generation Cuda Compute Architecture," Nvidia White Pap., 2012.