SOIL FERTILITY CLASSICATION VIA HYBRID HOPFIELD NEURAL NETWORK WITH ELECTION ALGORITHM FOR RANDOM BOOLEAN SATISFIABILITY REVERSE ANALYSIS
Keywords:
Hopfield neural network, Election algorithm, Reverse analysis, Random kSatisfiability, Soil fertility data setAbstract
Artificial neural network (ANN) possesses a comprehensive structure of training and testing stages that made it one of the most efficient tools in patterns and knowledge extraction in solving real-life applications such as classification, forecasting, risk analysis, detection and quantitative analysis. This paper presents a novel heuristic search and optimization-based algorithm inspired by political parties’ competition in the pursuit of power known as the Election Algorithm (EA). The purpose is to hybridize the Hopfield neural network (HNN) with a novel Election algorithm (EA) as a single neuro-dynamical model for high order random ksatisfiability reverse analyses in data mining application. The objectives are to reflect knowledge efficiently for the Soil fertility classification problem. The learning process of the Hopfield neural network (HNN) has undergone various modifications and improvements according to various optimization problems. However, the HNN model is associated with the drawback of being easily trapped in the local minimum solution. The Election algorithm (EA) is proposed to address these challenges to accelerate the learning phase of HNN for optimal classification problem. To prove the efficacy of the proposed hybridization, Soil fertility datasets (SFDS) has been employed in this study. The paper considers investigating the performance of algorithms on Hopfield neural network-based random Satisfiability Reverse Analysis (HNN-RANSATRA). The performance analysis of the proposed technique has been compared with the other two techniques in the Hopfield neural network-based Random Satisfiability Reverse Analysis with Artificially bee colony (HNN-RANSATRA-ABC) and Hopfield neural network-based Random Satisfiability Reverse Analysis with Exhaustive search techniques (HNN-RANSATRA-ES). The performance of all methods under study have been evaluated in terms of error accumulations and accuracy in the classification of Soil fertility data set (SFDS). The result of the experiment revealed that the proposed technique outperformed its two counterparts in terms of robustness, accuracy, and efficiency during the training process. Hence, based on the simulation results it has been proven that the Election algorithm complied effectively with the Hopfield neural network for Random Satisfiability Reverse Analysis for soil fertility data set (SFDS).
References
Abubakar, H. (2022). An optimal representation of optimal maximum random ksatisfiability in hopfield neural network for k3. Kuwait J. of Science., 49(2):1-16.
Abubakar, H. and Danrimi, M.L. (2021). Hopfield type of artificial neural network via election algorithm as heuristic search method for random boolean ksatisfiability. International Journal of Computing and Digital Systems.,10(1):659-673.
Abubakar, H., Masanawa, S. A., and Yusuf, S. (2020). Neuro-Symbolic Integration of Hopfield Neural Network for Optimal Maximum Random kSatisfiability (Maxrksat) Representation. Journal of Reliability and Statistical Studies., 13(1):199-220.
Abubakar, H., Muhammad, A., and Bello, S. (2022). Ants colony optimization algorithm in the Hopfield neural network for agricultural soil fertility reverse analysis. Iraqi Journal For Computer Science and Mathematics., 3(1):32-42.
Abubakar, H., Sabri, S. R. M., Masanawa, S. A., and Yusuf, S. (2020). Modified election algorithm in hopfield neural network for optimal random k satisfiability representation. International Journal for Simulation and Multidisciplinary Design Optimization., 11(1):16.
Abubakar, H., Sathasivam, S., and Alzaeemi, S.A. (2020). Effect of negative compaing strategies of election algorithm in solvingg optimzation problem. Journal of Quality Measurement and Analysis., 19(2):171-181.
Abubakar, H., Masanawa, A.S., Yusuf,S., and Boaku, G.I. (2021). Optimal representation to High order Random Boolean kSatisfiability via Election Algorithm as heuristics search approach in the Hopfied neural networks. Journal of the Nigerian Society of Physical Sciences., 3(3):201-208
Achlioptas, D. (2009). Random Satisfiability. Handbook of Satisfiability., 185:245-270.
Biere, A., Heule, M., and van Maaren, H. (Eds.). (2009). Handbook of satisfiability (185). IOS press.
Bhuyar, V. (2014). Comparative analysis of classification techniques on soil data to predict fertility rate for Aurangabad District. Int. J. Emerg. Trends Technol. Comput. Sci., 3(2):200-203.
Crestani, F. (1995). Neural networks for knowledge Representation and inference: DS Lavine and M. Aparicio IV (Eds.). Lawrence Erlbaum Associates Publishers, xv+ 503 99.95., ISBN 0-8058-1159-1.
Emami, H. (2019). Chaotic election algorithm. Computing and Informatics., 38(6):1,444-1,478.
Emami, H., and Derakhshan, F. (2015). Election algorithm: A new socio-politically inspired strategy. AI Communications., 28(3):591-603.
Emami, S., Choopan, Y., and Parsa, J. (2018). Modeling the Groundwater Level of the Miandoab Plain Using Artificial Neural Network Method and Election and Genetic Algorithms. Iranian journal of Ecohydrology., 5(4):1,175-1,189.
Feldman, V., Perkins, W., and Vempala, S. (2018). On the complexity of random satisfiability problems with planted solutions. SIAM Journal on Computing., 47(4):1,294-1,338.
Hopfield, J. J. (1982). Rigorous bounds on the storage capacity of the dilute Hopfield model. Proceedings of the National Academy of Sciences., 79:2,554-2,558.
Neelakanta, P.S. and De Groff, D.F. (2018). Neural network modeling: Statistical mechanics and cybernetic perspectives., CRC Press. 256.
Rojas, R. (1996). Threshold logic. In Neural Networks. Springer, Berlin, Heidelberg., 29-53.
Sathasivam, S. (2010). Upgrading logic programming in Hopfield network. Sains Malaysiana., 39(1):115-118.
Zubaidi, S.L., Abdulkareem, I.H., Hashim, K.S., Al-Bugharbee, H., Ridha, H.M., Gharghan, S.K., and Al-Khaddar, R. (2020). Hybridised artificial neural network model with slime mould algorithm: a novel methodology for prediction of urban stochastic water demand. Water., 12(10):2,692.
