This study explores how many people in Chennai, Tamil Nadu know about digital wallets as India moves towards a cashless economy. Surveying 300 individuals, we looked into their preferred e-wallets and reasons behind their choices. The results reveal that more people are using e-wallets for daily transactions and find them convenient. Notwithstanding, some users express reservations about the security and ease of use of ewallets. To enhance their services, we therefore recommend that e-wallet companies address these concerns. The study also finds varying levels of awareness of e-wallets among different age groups in Chennai, Tamil Nadu. The results from the Chi-square test revealed that the relationship between age group and the awareness of digital wallet is statistically significant.

Religion is a source of mercy and peace but misconceptions make some religious people become extremists. Today, social networks provide an abundant amount of data in many forms, texts, images, videos, etc… where people express themselves, their opinions, and affiliations on various topics. In this paper, we focus on analyzing comments and tweets on social networks and extracting text features. Therefore, we will carry out predefined algorithms for machine learning to analyze and classify a set of comments and tweets (text mining) from social networks. In this paper, we will work on two types of classification, the first is the religious classification (Muslim, Christian, Jewish, and Atheist) and the second is the extremism detection. The relationship between the religious text and the extremism text is not imperative; therefore, we will consider in this paper that the classification of the religious text is independent of the classification (detection) of the extremist text. We proposed specific features for discovering the classification. Our work use and compares a set of algorithms to classify comments in social networks. The proposed model has achieved 93.33% accuracy for religious classification and 97 % accuracy for extremism detection.

Virtual screening (VS) is a computer scheme used in the study of medicine development. VS is often used in computer-aided searches for novel lead compounds based on chemical similarity. Similarity retrieving is a technique for identifying molecules that are architecturally matched to a target chemical, which is beneficial in the discovery of new medicines. In the majority of traditional similarity methods, the molecular characteristics of biological and non-biologically linked activities are given equal weight. However, it has been shown that some distinguishing characteristics are more significant than others, depending on the chemical structure. As a result, this distinction should be considered when assigning a higher weight to each significant piece. The main objective for this study is to optimize weights of different similarity measures in data fusion for searching chemical database by applying a genetic algorithm (GA). In this paper, comparisons of various coefficient fusions were carried out. The results show that the Tanimoto, Cosine, Kulcznski (2) and Fossum coefficients are the best single coefficient. Cosine and Fossum coefficients gave the best combination for 2-coefficient fusion with weightings of 0.960 and 0.937, respectively. For 3-coefficient fusion, Russell-Rao, A Tanimoto and Cosine coefficient, of weightings 0.972, 0.960 and 0.960 respectively, give the best result. Combinations of Tanimoto and Cosine coefficients perform well and give a large number of actives. Using combination, with weights ranging between 0.0 and 1.0 generated by genetic algorithm, gave a better number of active than the non-weighted combination. Combining Cosine and Fossum coefficients without weights yields an average of 21.89% among the top 10% of compounds, whereas when a genetic algorithm (GA) is used to combine Cosine and Fossum coefficients with weights of 0.960 and 0.937, respectively, an average of 22.16% among the top 10% of compounds is obtained. Generally speaking, combinations of coefficients performed better than single coefficients.

With the emergence of Web 2.0, there is a burst of movie reviews on websites like IMDB, yahoo movies etc. These reviews contain public opinions which could help others to decide whether to watch it or not. The existing opinion mining systems perform the subjectivity detection of reviews at different levels: Document level, Sentence Level, and Feature level. In this paper, we identify the limitations of existing systems and provide contributory work on the issues identified for opinion mining from movie reviews. We compared our system with various thirteen popular opinion mining systems. We find that our method performs better in comparison to others.

Efforts to include many financially in Ghana continue to witness substantial progress as the nuptial between information communication technology (ICT) and finance thrives. However, except for ICT, the role of other actors in the development of the financial technology market (FinTech) of Ghana is unexplored. Therefore, the study explores the Ghanaian FinTech market development and financial inclusion using the Actor-Network Theory (ANT) and a multiple panel regression model. Specifically, we expound on the processes of building a variety of networks by embracing heterogeneous actors in the market. The focal actor MTN-Ghana through mobile phones interested others with the simplicity, ease of use, and accessibility of mobile money. Further, the study identified government, mobile phones, telecommunication companies, mobile money agents, mobile money customers, and banks as key actors in the market. Empirically, these actors show a significant positive relationship with financial inclusion. Therefore, these actors become significant to the existence, furtherance, and sustainability of the FinTech market and financial inclusion in Ghana.

The Colombian national policy for digital transformation and artificial intelligence aims to promote the generation of social and economic value in the country through the implementation and strategic use of new technologies in both the private and in the public sector, which allows improving the well-being of citizens and boosting productivity (CONPES, 2019). Being aware that Institutions of Higher Education are to lead digital transformation in their processes, this paper describes a digital transformation strategy applied in a Business Administration Program at a University in Colombia. The main objective of this project was to implement a virtual classroom for the management of degree projects of the Business Administration Program of a University in Colombia, as digital transformation strategy, in order to manage and control the activities of sending the project deliverables, qualification, evaluation, information and the course materials, facilitating communication between the different actors in this process, such as: research coordinators, methodological and disciplinary tutors, and students. An exploratory and applied research methodology was followed, using synchronous and asynchronous virtual education methods. For the implementation of the virtual classroom, the support of research methodology and disciplinary professors was required. The consultancies in the professional field were guided by professors specialized in ICT, through the use of virtual tools such as Moodle, Google Form for the completion of the evaluation rubrics by the juries and Microsoft Teams to record the online meetings carried out for the projects defense which allowed monitoring by the research coordination and the directives of the program. In conclusion, the implementation helped develop new skills and generate spaces both in experience and in innovation in terms of collaborative work within the Busniness Administration Program, leading this process through the digital transformation journey as well.

Although the exploration of Robotic Process Automation (RPA) has increased rapidly in the area of digital businesses, the benefits of using RPA in performing research activities (especially ANN-based research) have not been investigated thoroughly. Therefore, this work proposes and investigates a model approach for applying RPA in ANN-based research. This was conducted by forming an approach model of combining the automation principle of RPA in mimicking the manual tasks of ANN-based research of predicting the performance of high school students. Accordingly, the proposed model consists of the processes of (data gathering, data preprocessing, ANN implementation, and Result validation). Then the evaluation process in this work carried out by checking the time consumption and accuracy of using RPA method to automate the research process and compare it with manual process of conducting the same activities. The results show that using the proposed model of RPA in conducting ANN-based research activities significantly reduces time-consumption and increases result accuracy. The results will be of interest to the research community of RPA applications and ANN implementations.

Testing is a process of trying to find out every believable fault or weakness in a project. It is very crucial to sequence and writes optimized test cases to grow the rate of fault identification during the test design phase as early as possible. There are various proven test design techniques available, which focus on optimizing test sets in various test stages. However, system testing of complex applications often restricts access to the source code, i.e., they are a black box. Here, a great set of test cases is manually executed. In this article, a Deep Diverse Prototype Forest Model has been developed to prioritize the test cases. It will start with the collection of relevant test cases using the associate member function by Random Forest Algorithm. After that, the selected relevant test cases are given to the DDPF model to prioritize the test cases by minimizing the actual number of test cases. The implementation result shows that the proposed DDPF method improves the accuracy by 93.5% and reduces the execution time by 23% and 47% for the 2-objective and 3-objective models when associated with the present technique. It is concluded that the proposed DDPF method advances the detection of fault test cases accurately and eases the testing cost in the future.

In order to explore the determinants of vacant taxi driver\'s customer-search behavior, this paper intends to calibrate a time-dependent multinomial Logit (MNL) model by mining over 1.6 billion GPS records from about 8,400 taxis in Shanghai, China. First, based on the ordering points to identify the clustering structure (OPTICS) algorithm, the downtown area of the Shanghai city is divided into 47 hotspots to identify the hot areas of customer delivery and searching. Then, by investigating a typical search delivery process of a vacant taxi, five candidate factors that may affect the customer-search behavior are summarized and defined. Using the maximum likelihood method, the significant factors are finally found. The results reveal that the relative passenger demand, the regional likelihood of pick-ups as well as the expected rate of return are the most significant factors influencing the customer-search behavior. Although the impact of traffic situation (i.e., the en-route delay and traffic condition of the target hotspot) is not particularly significant, service providers and policymakers should still take full advantage of it to schedule taxi service and mitigate the traffic congestion caused by the circulation of vacant taxis. Besides, this paper also shows that the customer-search behavior of a vacant taxi driver varies with the time of day. Findings in this paper are expected to provide comprehensive insights about factors that should be considered in the future operation pattern of a taxi service system where human driver taxis and self-driving taxis are mixed.

QASs provide specific answers to the user’s questions. Most of the QASs focus their attention to factoid type questions [27]. Answers to such questions are found based on lexical and syntactical similarity between question text and source documents, in contrast, the complex questions such as WHY require semantic and higher level analysis of text to retrieve answers. In this paper, we make survey of research carried in WHY type QASs. We identify research issues and scope of improvements in WHY type QASs. \nIn this paper we survey literature on WHY type Question Answering Systems (QASs), and issue of sentiment analysis of opinion questions asked in a QAS. The paper investigates current state of the art of WHY type QASs with focus on the issue of computing sentiment polarity of opinion questions asked on product review sites.

Image segmentation (IS) is the most challenging and fundamental task in computer vision for analyzing medical images. Amongst several medical image segmentation methods available in the literature, Persistent Homology (PH) based methods are relatively new and have several good features with many applications in medical imaging. PH is a rapidly growing technique from Topological Data Analysis (TDA). TDA is a decade-old field built on the tools from Algebraic Topology (branch of pure mathematics), which studies the shape of data. TDA is rapidly growing due to its unique way of analyzing inferring exploiting complex data sets. In PH-based methods, images are converted to point clouds; then, simplicial complexes are constructed at different varying scales on the top of point clouds using the filtration process. Several topological features like connected components loops or holes are obtained during the filtration process at different scales. Then Image is segmented based on these features. This review is devoted to PH-based medical segmentation. All the methods are systematically reviewed. In addition, PH with a deep learning segmentation technique is also discussed.

Breast cancer ranks second, as a leading cause of cancer-related deaths in women. In fact, the annual incidence of breast cancer in the United States is 200,000. Early diagnosis of cancer creates a huge difference in prognosis and patient care. The current gold standard for detection of breast cancer is histopathological diagnosis- a tedious, and specialist dependent process that has a margin for error. Convolutional Neural Network (CNN) is a special type of deep feedforward neural network, which owing to image recognition and object classification provides a unique opportunity to create a more efficient and accurate diagnostic aid. The pixel data from the input image is organized in layers situated in all architectures but the difference between these architectures is the order and the parameters of the layers such as filter size, strides, padding etc. This study proposes utilizing deep learning method by applying GoogLeNet architecture for classification of Hematoxylin and Eosin (H&E) stained breast tissue images taken from BreaKHis and use the result after running GoogLeNet by three deferent learning rate optimization methods; root mean square propagation (RMSprop), adaptive moment estimation (ADAM) and stochastic gradient descent with momentum (SGDM). The histopathological images will be classified into either of two diagnostic classes i.e., benign or malignant. The architecture of GoogLeNet was modified for final two layers (fully connected layer, classification layer) to make it suitable for this study. The methodology was based on the utilization of transfer learning strategy and employing the GoogLeNet in classifying the full-database colored images after resizing them to be compatible with the selected CNN input. RMSprop was used as hyper-optimization method. This study for the proposed classification system for histopathological images of breast cancer from BreaKHis dataset utilizing GoogLeNet deep learning classification method and RMSprop as learning rate optimization method has yielded an accuracy of 96.27%, sensitivity of 90.43% and precision of 97.42%. This, by far, is the best result compared to other methods.

In this paper a Pareto multiobjective and Grasshopper optimization algorithm (GOA) based optimum Proportional–Integral-Derivative (P-I-D) controller design is proposed for improving the vehicle active suspension system dynamics under road disturbance conditions. The Pareto objectives considered are minimization of suspension deflection of sprung mass, tyre deflection, minimization of sprung mass acceleration and Eigen value based objective function. State space model for quarter vehicle active suspension system with P-I-D controller is developed for analyzing the stability and dynamic performance of the system. The sinusoidal based bump road disturbances are used for testing the robustness of the proposed control technique. Simulation results have been presented to show the advantage of proposed Pareto multiobjective objective and GOA based P-I-D controller over the weighted multiobjective objective and Genetic algorithm based P-I-D controller in terms of stability and dynamics of the active suspension system.

ABSTRACT\nPurpose: Evaluate the impact of ophthalmic lenses with a selective blue-violet light filter and an anti-reflexive coating with an additional power of +0.40 D in the near zone, on the convergence functions and in symptoms of digital asthenopia. Methods: A randomized study in cross-over design was conducted on 49 volunteers (age, 29 ± 5.5 years; male: female, 18:31). Each subject wore test lenses (+0.40 D in the near zone) and control lenses (regular single vision), for 4 weeks each in randomized order. Both lenses had a selective blue-violet light filter and an anti-reflective coating. A baseline measurement was taken with the subjects\' current updated glasses and then after wearing either the test lenses or control lenses. Accommodation amplitude (AA) and near point of convergence (NPC) were measured binocularly using the push-up technique with the RAF ruler. An average of three measurements was taken for the analysis. Digital asthenopia (DA) was evaluated by a questionnaire. Results: No significant difference (p=.52) was found for AA at baseline (11.50±1.88 D) and after 4 weeks wearing test lenses (11.61± 1.62 D) or control single vision (SV) lenses (11.88±1.50 D). Similarly, the changes in NPC, between baseline (6.50 ± 2.89cm) and measurement 4 weeks later wearing test lenses (6.71± 3.49) or control SV lenses (6.82± 3.50 cm) were statistically insignificant (p=.94). Comparison between the tested lenses and controls did not reveal significant differences in the values of the total asthenopia scores and between the mean scores of each of the symptoms of DA (p=0.20). Conclusions: The +0.40 D lenses have no negative impact on convergence, \"lazy accommodation\" or loss of accommodation power. The +0.40 D lenses and control SV lenses, both lenses with a selective blue‐violet light filter and an anti-reflexive coating, had a similar impact in attenuating the symptoms of DA

Developed intelligent technologies are become play a promising role in providing better decision-making and improving the medical services provided to the patients. A risk prediction task for short-term is big challenge task; however, it is a great importance for recommendation systems in health care field to provide patients with accurate and reliable recommendations. In this work, clustering method and least square support vector machine are used for prediction a short-term disease risk prediction. The clustering similar method is based on Euclidean Distance which used to identify the similar sliding windows. The proposed model is trained by using the slide windows samples. Finally, the appropriate recommendations are generated for heart diseases patients who need to take a medical test or not for following day using least square support vector machine. A real dataset which collected from heart diseases patient is used for evaluation. The proposed method yields a good results related by the recommendations accuracy generated to chronicle heart patients and reduce the risk of incorrect recommendations.