International Journal of Computational Intelligence and Applications

As real world data, larceny-theft rates are most likely to have both linear and nonlinear components. A single model such as the linear or nonlinear model may not be sufficient to model the larceny-theft rate. Thus, a hybridization of the linear and nonlinear models is proposed for modeling the larceny-theft rate. The proposed model combines Support Vector Regression (SVR) and Autoregressive Integrated Moving Average (ARIMA) models. Particle swarm optimization is used to optimize the parameters of SVR and ARIMA models. The proposed model is equipped with features selection that combines grey relational analysis and SVR to choose the significant economic indicators for the larceny-theft rate. The experimental results show that the proposed model has better accuracy than the linear, nonlinear, and existing hybrid models in modeling the larceny-theft rate of United States.

House fire is one of the major concerns for designers, builders, and residents of property. In the case of detecting fire, individual sensors have been used for a long time, but they cannot detect the level of fire and notify the emergency response units. To solve this problem, this study attempts to propose an intelligent early fire detection system that would not only detect the fire by using integrated sensors but also notify the appropriate authorities including fire department, ambulance services, and local police station simultaneously to protect valuable lives and properties. Signals from the integrated detectors e.g., heat, smoke, and flame go through the machine learning algorithms to check the potentiality of the fire as well as broadcast the predicted result to various parties using a GSM modem. To consolidate the predicted output, structured forest for fast edge detection has also been applied. The final outcome of this development also minimized false alarms, thus making this system more reliable.

The present contribution describes an application of a hybrid approach using response surface methodology (RSM) and particle swarm optimization (PSO) for optimizing the machining parameters in electric discharge machining (EDM) of compo casted Al6061/ cenosphere AMCs. Compo casting processing route was employed to prepare the AMCs. Each experimentation in EDM was performed under different machining conditions of Peak Current (Amp), Pulse on time ($\mu$s), and Flushing Pressure (bar). Performance characteristics such as material removal rates (MRRs), surface roughness (SR), and electrode wear rates (EWRs) were evaluated. A Taguchi L27 orthogonal array was considered to plan the experimentation and RSM was applied to model the inter relationship between the input process parameters and responses. A mathematic model has been developed to provide a fitness function to PSO by unifying the multiple responses. Finally, PSO was used to predict the optimal settings of the processing condition for the multi-performance optimization of the EDM operation. The experimental observations confirm the feasibility of the strategy and are in good accordance with the predicted value over a wide range of processing conditions employed in the process.

Reliability is deemed as an important issue in many broad applications, e.g., telecommunication systems, electric power systems, computation and parallel processing systems. In order to reach efficient performance and reliable/available design, we need to optimize the cost and reliability/availability of the proposed designs. The main contribution of this paper is to address the optimum model for repairable component with series-parallel structure using redundancy allocation problem (RAP). In this regard, two novel modified Ant Colony Optimization (ACO), are defined. In order to improve availability, ACOs employ visibility and pheromone update to optimize the RAP. The proposed methodology includes single objective to maximize the availability of case study system with two constraints-weight and cost of components. Hence, three meta-heuristic algorithms, i.e., ACO, Genetic algorithm (GA) and Particle Swarm Optimization (PSO) are applied to find the optimum structure. Finally, the simulation results of all the proposed meta-heuristic algorithms are compared. The comparison reveals that the modified ACO (MACO) provides the maximum availability among all the other meta-heuristic algorithms.

To overcome the poor population diversity and slow convergence rate of grey wolf optimizer (GWO), this paper introduces the elite opposition-based learning strategy and simplex method into GWO, and proposes a hybrid grey optimizer using elite opposition (EOGWO). The diversity of grey wolf population is increased and exploration ability is improved. The experiment results of 13 standard benchmark functions indicate that the proposed algorithm has strong global and local search ability, quick convergence rate and high accuracy. EOGWO is also effective and feasible in both low-dimensional and high-dimensional case. Compared to particle swarm optimization with chaotic search (CLSPSO), gravitational search algorithm (GSA), flower pollination algorithm (FPA), cuckoo search (CS) and bat algorithm (BA), the proposed algorithm shows a better optimization performance and robustness.

The Sentiment Analysis has been witnessing a booming interest in recent years, due to the enormous growth of digital content, and various types of online reviews such as product and movie reviews. The aim of Sentiment Analysis is to use automated tools to detect and classify subjective information from these reviews. Feature selection happens to be an important step to extract and select more efficient text features, and at the same time to try improve the performance of the used classifier for Opinion Classification task. This paper proposes a methodology based on Genetic Algorithms to optimize the feature selection process for polarity classification. First, it uses a supervised weighting method in order to prune the searching space then, this weighting method is combined with stochastic search methods that generate the next feature subset in a heuristic manner. In order to validate the proposed method, we compared it with three feature selection methods on different sizes of feature subsets. The experimental results show the efficiency of our proposed method.

Clustering is an important technology in data mining, which attempts to partition a set of objects into clusters based on the values of their attributes. $K$-means is a simple and efficient data clustering algorithm. However, it highly depends on the initial solution and is extremely easy to be trapped in local optima. In contrast, meta-heuristic algorithms show good performance to break through the local optima obstacle. In this paper, we propose an improved pigeon-inspired optimization (IPIO) algorithm towards resolving this problem. The algorithm uses an object-based initialization method to generate the initial population and introduces a parametric control strategy to navigate the flying direction. Meanwhile, the climb process of monkey algorithm (MA) with dimension by dimension improvement is adopted to strengthen the local search ability. In this paper, experiments over six real datasets are conducted to validate the effectiveness of IPIO. The experimental results show that IPIO is an efficient alternative in resolving the clustering analysis problem.