Showing papers in "Expert Systems With Applications in 2023"

Role of proliferation COVID-19 media chatter in predicting Indian stock market: Integrated framework of nonlinear feature transformation and advanced AI

Abstract: The outbreak of the COVID-19 pandemic has transpired the global media to gallop with reports and news on the novel Coronavirus. The intensity of the news chatter on various aspects of the pandemic, in conjunction with the sentiment of the same, accounts for the uncertainty of investors linked to financial markets. In this research, Artificial Intelligence (AI) driven frameworks have been propounded to gauge the proliferation of COVID-19 news towards Indian stock markets through the lens of predictive modelling. Two hybrid predictive frameworks, UMAP-LSTM and ISOMAP-GBR, have been constructed to accurately forecast the daily stock prices of 10 Indian companies of different industry verticals using several systematic media chatter indices related to the COVID-19 pandemic alongside several orthodox technical indicators and macroeconomic variables. The outcome of the rigorous predictive exercise rationalizes the utility of monitoring relevant media news worldwide and in India. Additional model interpretation using Explainable AI (XAI) methodologies indicates that a high quantum of overall media hype, media coverage, fake news, etc., leads to bearish market regimes.

The vehicle routing problem with simultaneous pickup and delivery and occasional drivers

Abstract: This research addresses the Vehicle Routing Problem with Simultaneous Pickup and Delivery and Occasional Drivers (VRPSPDOD), which is inspired from the importance of addressing product returns and the emerging notion of involving available crowds to perform pickup and delivery activities in exchange for some compensation. At the depot, a set of regular vehicles is available to deliver and/or pick up customers’ goods. A set of occasional drivers, each defined by their origin, destination, and flexibility, is also able to help serve the customers. The objective of VRPSPDOD is to minimize the total traveling cost of operating regular vehicles and total compensation paid to employed occasional drivers. We cast the problem into a mixed integer linear programming model and propose a simulated annealing (SA) heuristic with a mathematical programming-based construction heuristic to solve newly generated VRPSPDOD benchmark instances. The proposed SA incorporates a set of neighborhood operators specifically designed to address the existence of regular vehicles and occasional drivers. Extensive computational experiments show that the proposed SA obtains comparable results with the state-of-the-art algorithms for solving VRPSPD benchmark instances – i.e., the special case of VRPSPDOD – and outperforms the off-the-shelf exact solver – i.e., CPLEX – in terms of solution quality and computational time for solving VRPSPDOD benchmark instances. Lastly, sensitivity analyses are presented to understand the impact of various OD parameters on the objective value of VRPSPDOD and to derive insightful managerial insights.

A multi-stage melanoma recognition framework with deep residual neural network and hyperparameter optimization-based decision support in dermoscopy images

Abstract: This paper developed a novel melanoma diagnosis model from dermoscopy images using a novel hybrid model. Melanoma is the most dangerous and rarest type of skin cancer. It is seen because of the uncontrolled proliferation of melanocyte cells that give color to the skin. Dermoscopy is a critical auxiliary diagnostic method in the differentiation of pigmented moles, which show moles by magnifying 10–20 times from skin cancers. This paper proposes a multi-stage melanoma recognition framework with skin lesion images obtained from dermoscopy. This model developed a practical pre-processing approach that includes dilation and pooling layers to remove hair details and reveal details in dermoscopy images. A deep residual neural network was then utilized as the feature extractor for processed images. Additionally, the Relief algorithm selected practical and distinctive features from these features. Finally, these selected features were fed to the input of the support vector machine (SVM) classifier. In addition, the Bayesian optimization algorithm was used for the optimum parameter selection of the SVM method. The International Skin Imaging Collaboration (ISIC-2019 and ISIC-2020) datasets were used to test the performance of the proposed model. As a result, the proposed model produced approximately 99% accuracy for classifying melanoma or benign from skin lesion images. These results show that the proposed model can help physicians to automatically identify melanoma based on dermatological imaging.

Efficiency analysis technique with input and output satisficing approach based on Type-2 Neutrosophic Fuzzy Sets: A case study of container shipping companies

Abstract: This work tries to discuss and evaluate the advantages and superiorities of the extended Efficiency Analysis Technique with Input and Output Satisficing (EATWIOS) method based on Type-2 Neutrosophic Fuzzy Numbers (T2NFNs). The suggested model is maximally stable and robust by considering sensitivity analysis results which demonstrates a new performance analysis approach based on T2NFN sets. The proposed model deals with the input and output criteria and considers existing uncertainties arising from insufficient information and the dynamic structure of the industries. The model's basic algorithm has a unique structure compared to the previous performance analysis technique, and it does not require applying additional weighting techniques to identify the criteria weights. To the best of our knowledge, the extended version of the EATWIOS technique based on the T2NFN set is presented for the first time. The developed model provides reasonable and logical results to practitioners because it deals with satisfactory outputs instead of optimal outputs. This model is an immensely strengthened version of the EATWIOS technique, as the T2NFN sets treat predictable and unpredictable uncertainties. The suggested T2NFN-EATWIOS is then applied to a real-world assessment problem in the container shipping industry. The obtained results are pretty reasonable and logical. Moreover, the results of a comprehensive sensitivity analysis with three stages approve the robustness of the suggested model.

Developing a conceptual framework for short text categorization using hybrid CNN- LSTM based Caledonian crow optimization

Abstract: • To develop conceptual framework using four phases for short text categorization. • To propose hybrid convolutional CNN-LSTM for classifying the short text. • To utilize Caledonian crow optimization algorithm for optimizing the weight of CNN. • Getting attracted to the learning strategy of crows the NC 2 LO algorithm is employed. • Dataset based on online reviews is used to identify positive and negative reviews. Due to the continuous progression of social media networking sites, people share their thoughts, viewpoints, videos, speech and images through short texts. But the short texts are manually understandable but hard for the machine to collect data for clarification. The limited terms present in the short texts seem difficult while categorizing, analyzing as well as evaluating. Since social media contains a substantial amount of information, it is necessary to mine only useful information from the existing short texts. To achieve effectiveness and efficiency in short text categorization, the content-related characteristics derived from various machine learning techniques are admired. The significant objective of the proposed approach involves the categorization of short text and enhancing its accuracy. In this work, a hybrid convolutional neural network-long short term memory (CNN-LSTM) based new Caledonian crow optimization (NC 2 LO) model is utilized to classify the short texts. The crows utilize both social and asocial learning for developing tool modelling skills. Getting attracted to the behaviour of this variety of crows, and motivated by the learning strategy of crows, a new Caledonian crow optimization model referred to as NC 2 LO model is established. The conceptual framework of the proposed methodology consists of four different phases namely the data collection phase, data preparation phase, data pre-processing phase and short text categorization phase to classify the short texts with a high accuracy rate. Then, the proposed CNN-LSTM based NC 2 LO for short text categorization is evaluated using four different types of datasets namely IMDb, AG news, Twitter, and Tagmy News. Finally, the comparative analysis is carried out to evaluate the accuracy rate for various approaches and the analysis demonstrated that the proposed approach achieves a high accuracy rate of about 97%.

Fractal graph convolutional network with MLP-mixer based multi-path feature fusion for classification of histopathological images

Abstract: • A novel fractal GCN is proposed for classification of histopathological images. • FGCN effectively learn multi-level spatial features from histopathological images. • An MLP-mixer based multi-path feature fusion unit (MMFFU) is developed. • MMFFU effectively fuses multi-level graph representations in FGCN. The spatial information among different tissue components and multi-level features is important in histopathological images for pathologists to diagnose cancers. Graph convolutional network (GCN) can effectively learn these spatial features to improve the performance of computer-aided diagnosis (CAD) for histopathological images. The newly proposed GCN-based framework can effectively avoid the complex image preprocessing for graph construction, which integrates convolutional neural network (CNN) and GCN into a framework (named CNN-GCN). However, existing GCN generally cannot well learn multi-level graph features to further promote spatial feature representation. In this work, a new fractal GCN (FGCN) is proposed, which integrates the graph convolution into the fractal structure with different paths to learn multi-level graph representations. Moreover, a novel MLP-mixer-based Multi-path Feature Fusion Unit (MMFFU) is developed in FGCN to fuse these multi-level graph features from multiple paths. In MMFFU, the Mixer layer of MLP-mixer and non-local attention operation are designed to enhance the information communication of features, which further improves feature representation. The proposed FGCN is then embedded into the CNN-GCN framework (named CNN-FGCN) to perform the end-to-end classification of histopathological images. The experimental results on two public histopathological image datasets indicate the effectiveness of CNN-FGCN.

Multi-view dynamic graph convolution neural network for traffic flow prediction

Abstract: The rapid urbanization and continuous improvement of road traffic equipment result in massive daily production of traffic data. These data contain the long-term evolution of traffic flow and dynamic changes in the traffic road network. Due to the complex topology of the traffic road network, traffic flow prediction is challenging as it contains complex, multi-periodic patterns, and is often affected by sudden events. In this paper, we propose a Multi-View Dynamic Graph Convolution Network (MVDGCN) that captures different levels of spatial–temporal dependencies to predict traffic flow. Firstly, we use the coupling graph convolution network to learn the relationship matrix among stations dynamically, capturing the spatial dependencies at different levels in the traffic network. Secondly, we establish three encoder–decoders, representing hourly, daily, and weekly views, to extract the evolution law of traffic flow from three different time periods. Finally, we use the dynamic fusion module to merge the spatial–temporal dependencies extracted from the multi-view encoder–decoders. We conducted experiments on two real datasets, NYCTaxi and NYCBike, and found that our proposed MVDGCN model outperformed the best baseline, improving the RMSE, MAE, PCC, and MAPE by 12.9%, 6.2%, 0.8%, and 6.5% respectively on the NYCBike dataset and 9.2%, 4.2%, 4.6%, and 3.0% respectively on the NYCTaxi dataset. These results show that the proposed MVDGCN model performs better than state-of-the-art algorithms.

Scientometric review and analysis of recent approaches to stock market forecasting: Two decades survey

Abstract: Stock Market Forecasting (SMF) has become a spotlighted area and is receiving increasing attention due to the potential that investment returns can generate profound wealth. In the past, researchers have made significant efforts to forecast the stock market trends and predict the best time to buy, sell, or hold. The essence of past investigators’ various techniques and methods was to maximise the abundant opportunities that abound in the stock market trading and amass huge wealth from it. Over the years, no scientometric review has been conducted to scientifically map out the trends, progress, and limitations in the subject area. In this regard, this paper presents a pioneering scientometric review in SMF. It investigates a total of 220 reputable articles (2001–2021) to identify trends and patterns in stock market forecasting studies. VOSviewer software was used to conduct science mapping analysis. Actionable insights from the analysis explain significant metrics such as the top research outlets, most-cited articles, most co-occurred keywords, most influential countries, and much more. More so, a key finding in this paper is the introduction of a less computational approach that has the possibility of making a better forecast. Yet, past researchers have not thoroughly explored this option. This paper is beneficial to Early Stage Researchers (ESR), governments, funding bodies, managers, analysts, financial enthusiasts, practitioners, and investors, so as to understand the current progress and focus areas in stock market prediction.

Actual rating calculation of the zoom cloud meetings app using user reviews on google play store with sentiment annotation of BERT and hybridization of RNN and LSTM

Abstract: The recent outbreaks of the COVID-19 forced people to work from home. All the educational institutes run their academic activities online. The online meeting app the "Zoom Cloud Meeting" provides the most entire supports for this purpose. For providing proper functionalities require in this situation of online supports the developers need the frequent release of new versions of the application. Which makes the chances to have lots of bugs during the release of new versions. To fix those bugs introduce developer needs users' feedback based on the new release of the application. But most of the time the ratings and reviews are created contraposition between them because of the users' inadvertent in giving ratings and reviews. And it has been the main problem to fix those bugs using user ratings for software developers. For this reason, we conduct this average rating calculation process based on the sentiment of user reviews to help software developers. We use BERT-based sentiment annotation to create unbiased datasets and hybridize RNN with LSTM to find calculated ratings based on the unbiased reviews dataset. Out of four models trained on four different datasets, we found promising performance in two datasets containing a necessarily large amount of unbiased reviews. The results show that the reviews have more positive sentiments than the actual ratings. Our results found an average of 3.60 stars rating, where the actual average rating found in dataset is 3.08 stars. We use reviews of more than 250 apps from the Google Play app store. The results of our can provide more promising if we can use a large dataset only containing the reviews of the Zoom Cloud Meeting app.