Author
Mustafa Abdul
Bio: Mustafa Abdul is an academic researcher. The author has contributed to research in topics: Social media & Sentiment analysis. The author has an hindex of 1, co-authored 1 publications receiving 6 citations.
Topics: Social media, Sentiment analysis
Papers
More filters
••
TL;DR: The research uses a hybrid method of using Swarm Intelligence optimization algorithms with classifiers to classify a speaker's or a writer’s attitude towards various events or topics and arranging data into positive, negative or neutral categories.
Abstract: Recently, there are emergence and advent of data Inter-personal interaction web sites, micro blogs, wikis, in addition to Web applications and data, e.g. tweets and web-postings express views and opinions on different topics, issues and events in many applications, in addition to, different domains that includes business, economy, politics, sociology, and etc., which are resulted from offering immense opportunities for studying and analyzing human views and sentiment. The objective of sentiment analysis is to classify a speaker's or a writer’s attitude towards various events or topics and arranging data into positive, negative or neutral categories. Sentiment analysis means determining the views of a user from the textual content regarding that topic i.e. how one feels about it. It might be used to classify the text content. Various researchers have used a widespread sort of methods to teach the classifiers for the Twitter dataset with various results. The research uses a hybrid method of using Swarm Intelligence optimization algorithms with classifiers. For each tweet, pre-processing will be done by performing various processes i.e. Tokenization; removal of stop-words and emoticons;
7 citations
Cited by
More filters
••
TL;DR: The proposed sequential model is applied on two different datasets, Wisconsin diagnosis breast cancer dataset and Electronic Health Records (EHR), and it is shown that the proposed hybrid GWO-SVM model gives a fast convergence and achieves accuracy rate by 99.30%.
Abstract: Breast cancer is one of the most common types of cancer worldwide. Early detection of cancer increases the probability of recovery. This work has three contributions. The first contribution is improving the performance of support vector machine (SVM) using a recent grey wolf optimizer (GWO) for diagnosis breast cancer with efficient scaling techniques. The second contribution is proposing three efficient scaling techniques against the classical normalization technique. The last contribution is using a parallel technique which applies task distribution to improve the efficiency of GWO. The proposed sequential model is applied on two different datasets, Wisconsin diagnosis breast cancer (WDBC) dataset and Electronic Health Records (EHR). Experimental results of WDBC show that the proposed hybrid GWO-SVM model achieves 98.60% with normalization scaling. Also, using the proposed scaling techniques with the proposed GWO-SVM model gives a fast convergence and achieves accuracy rate by 99.30%. The parallel version of the proposed model achieves a speedup by 3.9 on four CPU cores. On the other hand, Experimental results of EHR show that the proposed hybrid GWO-SVM model achieves 93.26% with normalization scaling against 82.05 for SVM.
16 citations
••
TL;DR: Two hybrid machine learning models are proposed to predict the earthquake magnitude during fifteen days in the region of southern California and the simulation results showed that the FPA-LS-SVM model outperformed the F PA-ELM, LS-S VM, and ELM models in terms of prediction accuracy.
Abstract: This research proposes two earthquake prediction models using seismic indicators and hybrid machine learning techniques in the region of southern California. Seven seismic indicators were mathematically and statistically calculated depending on pervious recorded seismic events in the earthquake catalogue of that region. These indicators are namely, time taken during the occurrence of n seismic events (T), average magnitude of n events (M_mean), magnitude deficit that is the difference between the observed magnitude and expected one (ΔM), the curve slope for n events using inverse power law of Gutenberg Richter (b), mean square deviation for n events using inverse power law of Gutenberg Richter (η), the square root of the released energy during T time (DE1/2) and average time between events (µ). Two hybrid machine learning models are proposed to predict the earthquake magnitude during fifteen days. The first model is FPA-ELM, which is a hybrid of the flower pollination algorithm (FPA) and the extreme learning machine (ELM). The second is FPA-LS-SVM, which is a hybrid of FPA and the least square support vector machine (LS-SVM). These two models' performance is compared and assessed using four assessment criteria: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Symmetric Mean Absolute Percentage Error (SMAPE), and Percent Mean Relative Error (PMRE). The simulation results showed that the FPA-LS-SVM model outperformed the FPA-ELM, LS-SVM, and ELM models in terms of prediction accuracy.
8 citations
••
TL;DR: A hybrid approach that optimizes extreme learning machine (ELM) classifier with one of the most recent swarm intelligence algorithms which is grey wolf optimization algorithms (GWO).
Abstract: Sentiment analysis on social media is one of the most popular text mining application and many researchers have devoted more efforts in this interesting field. Sentiment analysis is a method for analyzing data and extracting the feeling it represents. Twitter is considered one of the most common social media forums used by people on various occasions to express their opinions and express feelings. Twitter's sentiment analysis has grown gradually over the past few decades. Due to the format of small tweet, a new dimension is created for problems such as slang usage, abbreviations, etc. This paper proposes a hybrid approach that optimizes extreme learning machine (ELM) classifier with one of the most recent swarm intelligence algorithms which is grey wolf optimization algorithms (GWO). GWO is used to self-adaptation of hidden neurons weights rather than manually selection. Also avoid the overfitting problem and make the model more generalized and robust. Results represented in this paper showed that the proposed hybrid model GWO-ELM overcame the problems found in classical ELM model and achieved best accuracy against all compared models.
6 citations
••
01 Jan 2020TL;DR: Object of the project is creating model for identifying provocative and fake comments, and two artificial neural networks are created: definition of sarcasm and definition of sentiment analysis.
Abstract: Nowadays internet-trolls have big impact on other users, it interferes with comfortable use. Objective of the project is creating model for identifying provocative and fake comments. Science articles about detection of trolls by hand were searched for making the criteria of relevant comments selection. To achieve the goal there were created two artificial neural networks: definition of sarcasm and definition of sentiment analysis. The program result is datasets of troll comments and fake comments, statistic and diagrams of definition.
4 citations
••
TL;DR: The decision tree algorithm is applied on the used breast cancer microarray dataset (BCMD), which contains 289 patients and 35981 attributes and principal components analysis (PCA) is used to reduce the number of attributes.
Abstract: Badr et al. [1] proposed efficient scaling techniques EST with support vector machine on the data set Wisconsin from UCI machine learning with a total 569 rows and 33 columns. In this work, we try to evaluate the validity of the results reached by Badr et al. [1] in the case of using different datasets, different classifiers and dimensionality reduction tools? So, the decision tree algorithm is applied on the used breast cancer microarray dataset (BCMD) contains 289 patients and 35981 attributes. We use principal components analysis (PCA) to reduce the number of attributes. We also propose new scaling techniques to improve the accuracy of the decision tree algorithm. Experimental results show that the decision tree algorithm with new scaling techniques (equilibration, geometric mean and arithmetic mean) achieves 84.98 %, 80.65 % and 79.96 % accuracy against to the traditional normalization (normalization [0, 1], normalization [-1, 1] and standard normalization) by 75.44 %, 76.85% and 78.93%. General Terms Data Mining, Classification
2 citations