Narrow Results

In this paper, we propose a method to reduce the effect of light intensity on the extraction of skin information from real facial images. We have previously studied methods for estimating autonomic nervous activity, which is an indicator of stress, using real facial images. The results showed the possibility of assessing stress by capturing peripheral blood flow changes from the R–B values in the nasal area. However, it is possible that reflections and shadows caused by changes in the amount of light hitting the skin due to body movements, such as facial movements, could become noise. Therefore, this study proposes a weighting method to reduce the effect of light intensity caused by body movements quantitatively. We believe that this method can establish a stress evaluation method independent of the shooting environment. The results of the validation experiments using the proposed method suggest that it is possible to quantitatively reduce the effect of light intensity and evaluate peripheral blood flow conditions from actual facial images with high accuracy. This indicates the usefulness of the proposed method for stress evaluation.

Along with the developments of numerous MaOO algorithms in the last decades, comparing the performance of MaOO algorithms with one another is also highly needed. Many studies have attempted to manipulate such comparison to analyze the performance quality of MaOO. In such cases, the weight of importance is critical for evaluating the performance of MaOO algorithms. All evaluation studies for MaOO algorithms have ignored to assign such weight for the target criteria during evaluation process, which plays a key role in the final decision results. Therefore, the weight value of each criterion must be determined to guarantee the accuracy of results in the evaluation process. Multicriteria decision-making (MCDM) methods are extremely preferred in solving weighting issues in the evaluation process of MaOO algorithms. Several studies in MCDM have proposed competitive weighting methods. However, these methods suffer from inconsistency issues arising from the high subjectivity of pairwise comparison. The inconsistency rate increases in an exorbitant manner when the number of criteria increases, and the final results are affected. The primary objective of this study is to propose a new method, called a Novel Fuzzy-Weighted Zero-Inconsistency (FWZIC) Method which can determine the weight coefficients of criteria with zero consistency. This method depends on differences in the preference of experts per criterion to compute its significance level in the decision-making process. The proposed FWZIC method comprises five phases for determining the weights of the evaluation criteria: (1) the set of evaluation criteria is explored and defined, (2) the structured expert judgement (SEJ) is used, (3) the expert decision matrix (EDM) is built on the basis of the crossover of criteria and SEJ, (4) a fuzzy membership function is applied to the result of the EDM and (5) the final values of the weight coefficients of the evaluation criteria are computed. The proposed method is applied to the evaluation criteria of MaOO competitive algorithms. The case study consists of more than 50 items distributed amongst the major criteria, subcriteria and indicators. The significant contribution of each item to the algorithm evaluation is determined. Results show that the criteria, subcriteria and their related indicators are weighted without inconsistency. The findings clearly show that the FWZIC method can deal with the inconsistency issue and provide accurate weight values to each criterion.

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.