Data Science and Interdisciplinary Research: Recent Trends and Applications

Rainfall is one of the most considerable natural occurrences, which is important for both human beings and living beings. Since the environment is changing and there is a huge change in weather, it is noted that the rainfall cycles are also varying and the earth’s temperature is increasing day-by-day. The changes in weather conditions like humidity, pressure, wind speed, dew point and temperature affect the agriculture, industry, production, and construction and also lead to floods and land-slides. Hence it is one of the important factors to be noted for human beings to keep track of the natural occurrences in order to survive. In order to overcome these issues, a system is required which is able to forecast and predict the rainfall using statistical techniques which is the most popular tool in modern technology. This paper provides a detailed survey and comparative analysis of various methodologies used in the prediction of rainfall over multiple countries. Comparison is made in terms of various performance measures: accuracy, precision, recall, RMSE, specificity, sensitivity, MAE, F-Measure, ROC and RAE. Further, the drawbacks with existing approaches applied so far in the prediction are discussed. 

In the present situation, with the enhancement of virtualization techniques, it is very essential to keep track of accumulated large-scaled heterogeneous data in every respect. In addition to that, it is also necessary to prioritize the processing mechanisms when being linked with clustered data. Sometimes it has been observed that the large scaled datasets are too complex and therefore, the normal computation mechanisms are not sufficient or adequate for the specific applications. But it is highly required to observe the significance of each individual dataset and focus on the responses being accumulated from other aspects to make a suitable decision and generation of efficient analytical clustered data. The main aim of such applications is to apply the clustering gaining merits from evolutionary computation to process the large-scaled data and based on optimality, the performance of datasets can be measured. 

This chapter describes the initial working plan and the project carried out, along with their respective analyses. The project focused on implementing a cryptographic solution in the embedded system to ensure the authentication and integrity of CAN messages. With the rise of the autonomous driving concept in the automotive industry, there is a growing need for digital data processing and communication both within the vehicle and with the external world, creating a significant challenge to protect the data from cyber-attacks during communication. Hence, cybersecurity has become an important topic in the automotive industry. This project was carried out to ensure the security of data during communication on the vehicle Controller Area Network (CAN) using message authentication. The chapter concludes by highlighting the current technology's drawbacks and discussing potential future improvements.

 Agriculture is the backbone of an Agro-based Country's Economic System as it employs the majority of the population. Internet-of-Things (IoT)-based intelligent systems help reduce losses and make efficient use of available resources. This paper aims to detect anomaly conditions that might occur in sensor nodes related to day-t- -day smart irrigational activities in an agricultural field. IoT-based irrigation systems being prone to unauthorized intrusion can cause damage to smart farms in terms of crop damage and infertility of the soil. In this paper, we propose an intelligent decision-making system that can identify Anomalous Conditions and Suspicious Activities. The model discussed in this paper uses the idea of Gaussian distribution, which calculates the expected probability of a given state of an agricultural field and classifies anomalies based on what previous probabilities of an anomaly state looked like. The approach classifies the anomalies with an accuracy of 80.79%, a precision of 0.81, and a recall of 0.54 under test conditions.

Growing electricity needs among the vast majority of the population seconded by a voluminous increase in electrical appliances have led to a huge surge in electric power demands. With thediminishing unit price of electric meters and increase of loading, it has been observed that a certain amount of electric meters generate faulty readings after exhaustive usage. This results in erroneous meter readings thereby affecting the billings. We propose a fault detecting learning algorithm that is trained by early meter readings and compares the actual meter reading (AMR) with the predicted meter reading (PMR). The decision matrix generates an alarm if |PMR-AMR|>T; where T equals the threshold limit. T itself is decided by the learning algorithm depending upon the meter variance. Moreover, our system also detects if there is any power theft as such an action would result in a sudden rise in AMR. The learning algorithm deploys six binary classifiers which reflect an accuracy of 98.24% for the detection module and an error rate of 1.26% for the prediction module. 

Technology is a fundamental part of the teaching-learning process which has brought plenty of benefits during the pandemic situation. Covid 19 has adversely affected the education system throughout the world. Due to this abrupt change, mobile learning has occupied a dominant place in helping students to handle this unavoidable crisis. The studies focused on online learning, e-learning and also m-learning. However, this study focuses on the effectiveness of m-learning by understanding the satisfaction and intention of the students towards m-learning. Mixed research method approach is used in this study. Structural Equation Modelling (SEM) is used to validate the proposed model. The findings revealed that the majority of the students are satisfied with m-learning. However, there are also some students who express dissatisfaction with learning practical courses through mobile phones, citing insufficient support from their parents and the institution. Additionally, the satisfaction and intention of students play a crucial role in determining the effectiveness of mobile learning (m-learning). The study presents a comprehensive framework for understanding, explaining, and predicting the factors that influence the effectiveness of mobile learning (m-learning) among higher education students. The study also supports the practitioners and educators with useful guidelines for designing a successful m-learning system, particularly in higher education. This will also enable government to frame its the digitized policies appropriately. 

Many of the country's most important essential zones and urban regions are found on the outskirts of the coastal zone. The activity or processes of socio-economic growth among coastal areas on the usage of coastal resources are essential for understanding the socio-economic state of a region and its long-term management. Sustainable development shows two great notions for the protection of natural resources: (i) environmental protection and (ii) economic development. Development activities are allowed as long as they do not compromise people's quality of life or the viability of the natural systems on which development is based, according to the above two notions. Advanced scientific and technological knowledge is now helping to improve and prevent the unsustainable exploitation of natural resources along the Indian coast. River floods, cyclones, depressions, and, most crucially, coastal erosions are common occurrences in the studied region. This effect can potentially change the socio-economic situations of coastal residents and society. As a result of this background, the author decided to quickly analyze socio-economic conditions and look at the current state and impacts on the study area's socio-economic conditions. Information on the above parameters was gathered to get insight into the research area's socio-economic profile. The current project involves conducting a socio-economic survey on Orissa's southern coast. 

Electrocardiogram (ECG) records cardiac electrical signal to check for various heart problems. However, it can be impaired by noise. Therefore, ECG signal denoising is a significant pre-processing step that reduces noise and emphasizes the characteristic waves in the ECG data. The frequency range of a simple ECG is usually between 0.5 Hz and 100 Hz. When processing the ECG signal, artifact elimination is the most important resource since artifacts in ECG signal impede the diagnosis of disorders. This work uses MATLAB to reduce noise by applying low pass, high pass, and derivative pass filters. On the PTB database, the performance of these approaches is compared using benchmark measures such as mean-square error (MSE) and signal-to-noise ratio (SNR) to compare various ECG denoising algorithms. The combination of low pass + high pass + derivative pass filters produces low mean-square error (MSE) and signal-to-noise ratio (SNR) values of 0.052 db and 1.185 db when compared to the raw signal.

 In the biomedical research areas of whole genome sequence (WGS) analysis, disease diagnosis, and medication discovery, Next Generation Sequencing (NGS) data are the most recent and popular trend. The use of NGS data has improved the analysis of infectious diseases, WGS, illness identification, and medication discovery. Although the amount of NGS data is massive, researchers have worked and are continuously working to improve its quality and precision. Modern computational techniques increase the biological value of NGS data processing, making it more accessible to biomedical researchers. Although the complexity of NGS and the required computational power to analyse the data pose a significant threat to researchers, the introduction of various branches of Artificial Intelligence (AI) such as Machine Learning (ML) and Deep Learning (DL) has given analysis, prediction, and diagnosis a new direction. Deep Learning's potential has been demonstrated in a variety of fields, including biomedical research, where it has outperformed traditional methods. The development of deep learning algorithms aids in the analysis of complicated datasets such as NGS by giving a variety of advanced computational methodologies. Different DL approaches are designed to manage enormous datasets and multiple jobs, and the genetic research business could be the next industry to benefit from DL. This paper discusses a variety of DL methods and tools for analysing NGS data in the fields of contagious diseases, WGS analysis, disease diagnosis, and drug design.

Machine learning is applied in medical diagnosis to do early prediction of diseases, for increasing the possibility of recoverability around the globe. Cancer is a disease, which spreads quickly and would be difficult to control in advanced stages. The idea is to diagnose the disease at an early stage, so as to increase the chances of fast recovery. Breast cancer is common in women, and is a disease that causes the death of women in the age of fifty years or older. The purpose is to apply machine learning concepts to do early detection of disease. The system is fed with the images of all stages of cancer patients and the classification tools are used to train the system with the cases. This helps to predict the stage of cancer. After the prediction of the stage, the patient is prescribed with the medication or other appropriate treatment processes by the doctor. The right time diagnoses help to improve the prognosis and increase the chances of survival. The type of the tumour, size and its re-occurring nature need to be monitored from time to time to check it in control. The Data Mining algorithm in collaboration with Deep learning or Machine learning concepts can be used to design a system for early predictions. The proposal is to use the machine learning concepts to do performance comparison using different classifiers, such as Support Vector Machine (SVM), Decision Tree and K-Nearest Neighbour (KNN) on the Wisconsin Diagnostic Breast Cancer (WDBC) dataset [1]. The main aim of cancer detection is to classify tumours into malignant or benign, thus we use machine learning techniques to improve the accuracy of diagnosis. The main objective is to assess the efficiency, effectiveness and correctness of the algorithm using performance metrics like Accuracy, Precision, F1 score and Recall Experimentation is done using Jupyter Notebook.