MIT-World Peace University (MIT-WPU) is a top educational institute located in Pune, India. It is recognized within&nbsp;UGC under Government of Maharashtra Act No. XXXV 2017.&nbsp;

It is part of the MIT group of institutions (established in 1983) which covers 10+ campuses across India and has more than 50,000+ students enrolled across various disciplines at a given point of time. The institute houses 15 schools ranging from Engineering &amp; Technology, Design, Pharmacy, Commerce, Governance among others. It offers undergraduate, postgraduate and PhD programs among different disciplines.&nbsp;

&nbsp;

MIT World Peace University

School of Computer Engineering & Technology

In Telecom Industry customer churn is a big issue and one that impacts their revenue. When customers start to leave a service or subscription, it increases the expenditure for these companies. Businesses have found that acquiring new customers costs them nearly six times more money than retaining existing ones. Therefore, preventing customer churn becomes important when companies are trying to grow their business. The analysis of Customer Behaviour using Machine Learning techniques does provide an effective solution to the problem by predicting which customers are more likely to leave the service or subscription. Predictive analysis of customer behavior not only helps companies fix issues with their service but also helps them add new features and products so as to keep the customer engaged. The present work provides an overview of the latest works in the field of Customer Churn prediction. Our aim is to provide a simple path to make the future development of novel Churn prediction approaches easier.

Engineering and Technology

Journal of Emerging Technologies and Innovative Research

An Overview of Customer Churn Prediction in Telecom Industry

Sign Language is the primary mode of interaction between the specially-abled deaf and mute population. As per the World Health Organization (WHO) survey, 460 million individuals including men, women and children suffer hearing disability. Out of these 460 million individuals, about 12.3 million belong to India. Most of this population relies on a third person who is capable of translating hand sign gestures to the regionally spoken language for them. But involving a translator hinders the conversation privacy. Here, the rapid development in the field of Artificial Intelligence (AI) and Machine Learning (ML) aids to convert hand sign gestures into words that can be understood by normal people. The Indian sign language interpreter for the mute and deafened people follows a Vision-based approach which uses the Machine Learning technique of 3D-CNN (3-Dimensional Convolutional Neural Network) and LSTM (Long Short-Term Memory) neural network to productively map the input-Indian sign language gesture images to their meaning. This system aims to interpret hand gestures images of Indian Sign Language for alphabets, numbers, sentences as well as emergency words (conversion of images of hand gesture to text/audio). This proposed technique with data from various sources will aid to alleviate the communication gap between the deaf and mute people and others.

International Journal of Creative Research Thoughts

Indian Sign Language Interpreter for Deaf &amp; Mute People

According to the dual inheritance theory, human behavior is a result of genetic as well as cultural evolution. The socio- cultural inspired algorithms are developed by observing human behavior in a society. These social interactions contribute to the progress of humanity. This class of algorithms simulates various social setups like an election, a sports championship, an imperialistic colony, a family, or a group of friends who help other members in the group to improve their behavior while fulfilling their own goals. Some setups show a cooperative nature while others exhibit the competitive nature of humans. Like other Evolutionary algorithms, Social algorithms too are based on the principle of ‘survival of the fittest’ and are metaheuristics used for optimization problems. This work is aimed to give a brief overview of 15 Social algorithms with their advantages, limitations, and applications solved, so as to provide a succinct, yet comprehensive, outline of these techniques.

Society-inspired Evolutionary Algorithms: A Compendium

Society-inspired Evolutionary Algorithms: A Compendium

Human Activity Recognition (HAR) has become one of the most important research fields to achieve real-time monitoring of human activities for timely decision making in various applications like fall detection, elderly care etc. Now-a-days, most people use smartphones which come with various embedded inertial sensors like accelerometer and gyroscope to monitor acceleration and angular velocity. These smartphone-based sensors have proven to be cost-effective solution in identification of activities belonging to ADL (Activities of Daily Living). Various Machine Learning, Deep learning and hybrid models have been proposed and implemented for HAR. This paper also proposes various hybrid deep learning approaches which combine Deep Neural Networks with other models like LSTM (Long Short Term Memory) Model and GRU (Gated Recurrent Unit) for effective classification of engineered features from CNN (Convolutional Neural Network) Model. A novel architecture that integrates CNN with Random Forest Classifier (DeepCNN-RF) is proposed to add randomness to the model. The proposed models have been tested on publicly available HAR Datasets like UCI HAR and WISDM Activity Recognition Datasets. Experimental results show that the hybrid models outperform the state-of-the-art data mining, machine learning techniques in UCI HAR and WISDM with an overall maximum accuracy of 97.77% and 98.2% respectively.

World University root node

Multimedia Tools and Applications

Hybrid deep learning approaches for smartphone sensor-based human activity recognition

Query processing can be briefly defined as a database that comprises of an organized collection of data for one or more users either in digital form or in analog form such that it can portray exactly. A Wireless Sensor Network (WSN) is a specialized network of minimum cost and power sensor nodes that can be described as the ability of performing some processing, gathering sensory information and communicating with each other. Query ordering with data aggregation is the process of scheduling of the nodes to receive the useful data from sensors. Data aggregation is considered as one of the fundamental processing procedures for saving the energy. In WSN data aggregation is an effective way to save the limited resources. This paper proposes a novel query-based data aggregation model with the aid of intelligent techniques. The framing of the query order takes place and the frames are ranked on the basis of a multi-objective function. The newly developed multi-objective function includes Latency, Throughput and Data freshness. Initially, the solution corresponding to query order is trained in NN using the proposed Fitness-Mated Lion Algorithm (FM-LA). The optimally generated query order from NN is further given for second-level solution generation, which is again applied to FM-LA for subsequent query order optimization. Hence the two-stage optimization process with NN for query ordering is compared over the conventional methods in terms of performance measures like Latency, throughput and data freshness. Hence, substantiated performance and comparative analysis validate the improved performance of the proposed model.

Wireless Personal Communications

Combination of Fitness-Mated Lion Algorithm with Neural Network for Optimal Query Ordering Data Aggregation Model in WSN

Cybernetics, Cognition and Machine Learning Applications

Prediction of Stock Indices, Gold Index and Real Estate Index Using Deep Neural Networks

Supervised algorithms depend on the given data for categorizing. In present work, we used both parametric and nonparametric types of classifiers. We intend to compare the performance of four popular machine learning classification algorithms—Naïve Bayes, decision trees, logistic regression, and random forest on two popular benchmarked datasets—wine quality dataset and glass identification dataset. To get a wide angle of the performance of these algorithms, we incorporated both binary and multi-class classification which also solved the problem of imbalance in the dataset. In current work, we compare and demonstrate various supervised machine learning classification algorithms on the two well-known datasets. The performance of the algorithms was measured using accuracy, recall, precision, and F1-score. It was observed that nonparametric algorithms like random forest classifier and decision tree classifier bested the parametric algorithms like logistic regression and naïve Bayes. Moreover, as the datasets were imbalanced, we figured out which algorithm performs better under what circumstances. In particular, random forest achieved best performance in terms of all considered metrics, with accuracy of 82 and 83% in wine datasets and 79% in glass identification dataset. © 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Lecture Notes on Data Engineering and Communications Technologies

Investigating the Impact of Data Analysis and Classification on Parametric and Nonparametric Machine Learning Techniques: A Proof of Concept

Determining the image sentiment is a tedious task for classification algorithms, owing to complexities in the raw images as well as the intangible nature of human sentiments. Classifying image sentiments is an evergreen research area, especially in social data analytics. In current times, it is a common practice for majority people to precise their feelings on the web by substituting text with the upload of images via a multiplicity of social media sites like Facebook, Instagram, Twitter as well as any other platform. To identify the emotions from visual cues, some visual features as well as image processing techniques are used. Several existing systems have already introduced emotion detection using machine learning techniques, but the traditional feature extraction strategies do not achieve the required accuracy on random objects. In the entire process, normalization of image, feature extraction, and feature selection are important tasks in the train module. This work articulates the newest developments in the field of image sentiment employing deep learning techniques. Also, the use of conventional machine learning techniques is compared along with deep learning algorithms. It is indicative that a combination of fast recurrent neural networks and CNN may produce high accuracy with minimum time complexity. It is noted from the survey that existing researchers reflect CNN provides around 96.50% average accuracy for sentiment classification on the flicker image dataset. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd 2021.