The huge text generated on social media in Arabic, especially the Arabic dialect becomes more attractive for Natural Language Processing (NLP) to extract useful and structured information that benefits many domains. The more challenging point is that this content is mostly written in an Arabic dialect with a big data fusion challenge, and the problem with these dialects it has no written rules like Modern Standard Arabic (MSA) or traditional Arabic, and it is changing slowly but unexpectedly. One of the ways to benefit from this huge data fusion is opinion mining, so we introduce this systematic review for opinion mining from Arabic text dialect for the years from 2016 until 2019. We have found that Saudi, Egyptian, Algerian, and Jordanian are the most studied dialects even if it is still under development and need a bit more effort, nevertheless, dialects like Mauritanian, Yemeni, Libyan, and somalin have not been studied in this period. Many data fusion models that show a good result is the last four years have been discussed.
Read MoreDoi: https://doi.org/10.54216/FPA.090101
Vol. 9 Issue. 1 PP. 08-28, (2022)
One of the important schemes for modern communication is Filter Bank Multi-Carrier with Offset Quadrature Amplitude Modulation (FBMC/ OQAM), as it provides better spectral efficiency with small inter-symbol and inter-carrier interference specially in data fusion platforms. Unfortunately, the design of filter banks in FBMC is difficult and complex to achieve the requirement due to complexity of handling the data fusion issues. This paper presents a proposed method to design a uniform and nonuniform filter banks using a data fusion optimization technique. The design process represented as an objective function describes the amplitude in the stop band, and the goal is to minimize the objective function. Different examples are provided to illustrate the efficiency of the proposed design method.
Read MoreDoi: https://doi.org/10.54216/FPA.090102
Vol. 9 Issue. 1 PP. 29-37, (2022)
The last decade lead to an unbelievable growth of the importance of social media. Due to the huge amounts of documents appearing in social media, there is an enormous need for the automatic analysis of such documents. In this work, we proposed various regression models for the blog feedback prediction to be used in the data fusion environment. These models include decision tree regressor, MLP regressor, SVR, random forest regressor, and K-Neighbors regressor. The models are enhanced by average ensemble and ensemble using K-Neighbors regressor. The Blog Feedback dataset is used for training and evaluating the proposed models. The results show that there is a decrease in RMSE, MAE, MBE, R, R2, RRMSE, NSE, and WI when compared to the traditional methods.
Read MoreDoi: https://doi.org/10.54216/FPA.090103
Vol. 9 Issue. 1 PP. 38-46, (2022)
Automatic classification of biomedical signals helps to perform decision making in the healthcare sector. Electrocardiogram (ECG) is a commonly employed 1-dimensional biomedical signal that can be utilized for the detection and classification of cardiovascular diseases. The recently developed deep learning (DL) models find useful for the detection and classification of ECG signals for cardiovascular diseases. With this motivation, this study develops an intelligent electrocardiogram classification using sailfish optimization algorithm with gated recurrent unit (SFOA-GRU) technique. The goal of the SFOA-GRU model is to detect the existence of cardiovascular disease by the classification of ECG signals. The SFOA-GRU model initially undergoes data pre-processing step to transform the actual values into useful format. Besides, GRU model is applied for the detection and classification of ECG signals. For improving the classification outcomes of the GRU model, the SFOA has been utilized to optimally adjust the hyper parameters involved in it. A wide-ranging experimental analysis is carried out to demonstrate the enhanced outcomes of the SFOA-GRU model. A comprehensive comparative study highlighted the promising performance of the SFOA-GRU model over the other recent approaches using different measures parameters.
Read MoreDoi: https://doi.org/10.54216/FPA.090104
Vol. 9 Issue. 1 PP. 47-58, (2022)
Stock exchanges are developed as an essential component of economies, as they can promote financial and capital gain. The stock market is network of economic connections where share is bought and sold. Stock Market Prediction (SMP) is quite useful to investors. An effective forecast of stock prices is offer shareholders with suitable help in making appropriate decisions regarding if sell or purchase shares. The employ of Machine Learning (ML) and Sentiment Analysis (SA) on data in microblogging sites are developed as a famous approach to SMP. However, the heterogenous data fusion in stock market field is a big challenge. This paper introduces an effective Cat Swarm Optimization with Machine Learning Enabled Microblogging Sentiment Analysis for Stock Price Prediction technique. The presented model investigates the social media sentiments to foresee SPP. Firstly, the proposed model executes data pre-processing and Glove word embedding approach. Next, the weighted extreme learning machine approach was utilized for the classification of sentiments for SPP. Lastly, the CSO system was exploited for optimal adjustment of the parameters related to the WELM model. The experimental validation of the proposed approach was executed using microblogging data. The results show that the proposed method outperforms the previous studies.
Read MoreDoi: https://doi.org/10.54216/FPA.090105
Vol. 9 Issue. 1 PP. 59-69, (2022)
Leukemia is a type of blood cancer. Leukemia is cancer that begins in the blood cells. The lymphocytes and other blood cells are created in the bone marrow. When a person has leukemia the bone marrow does not function properly. Leukemia cells are produced by the bone marrow. Leukemia cells are mainly referred to as "rupture". These naive cancer cells block the cells that create the bone marrow. In this paper, various approaches to the classification & automatic detection of leukemia are described. The experiment was successfully implemented in Kaggle. Deep Learning algorithms were largely used in the treatment of Leukemia for the classification & detection of its presence in a patient. The paper describes Convolutional Neural Networks (CNN) and Visual Geometry Group-16(VGG-16) algorithms that are used to categorize leukemia into its sub-types and presents a comprehensive study of these algorithms.
Read MoreDoi: https://doi.org/10.54216/FPA.090106
Vol. 9 Issue. 1 PP. 70-76, (2022)