Journal of Innovations in Data Science and Big Data Management (e-ISSN: 2583-9845)

Hand Gesture Recognition in Real-Time Using Deep Learning Scheme

S. Karthikeyini, M. Rupa, Ravikumar. M, Athira.S — Tue, 31 Oct 2023 00:00:00 +0530

Deep learning algorithms are employed in
gesture recognition because of their capacity
to autonomously acquire and extract features
from raw data, rendering them highly
suitable for intricate pattern identification.
Through the application of data acquisition
and feature extraction techniques,
information about hand characteristics and
the distinctive components of gesture
movements are obtained. The gesture
recognition algorithm is then applied to
evaluate the dataset and learn to discriminate
between various hand positions. The
proposed work deep learning includes a
Convolutional Neural Network used to
recognize the gestures to provide immediate
feedback to a user. Hand gesture recognition
can be challenging due to variations in
lighting, background clutter, occlusions, and
differences in hand size and shape. Robust
algorithms are used to account for these
challenges. CNNs are not limited to image
processing alone. They have found
applications in natural language processing,
medical image analysis, and even gaming.
Their versatility and adaptability have made
them a go-to choice for tasks that involve
complex data patterns.

A Critical Study on the Application of Explainable AI for Handling Data Poisoning Attacks

Manas Kumar Yogi — Thu, 23 Nov 2023 00:00:00 +0530

With the rapid integration of machine
learning models into various critical
applications, the susceptibility of these models
to adversarial attacks has emerged as a
significant concern. Data poisoning attacks, a
subset of adversarial attacks, involve injecting
malicious or misleading data into the training
set with the intent to degrade the model's
performance or induce erroneous predictions.
Explainable Artificial Intelligence (XAI)
techniques have gained prominence as a
promising approach to enhance the resilience
of machine learning models against such
attacks. This paper explores the role of
Explainable AI in mitigating data poisoning
attacks by providing interpretable insights
into the model's decision-making process. We
review various XAI methods that aid in
detecting and mitigating the presence of
poisoned data during both the training and
inference phases. Additionally, we discuss how
XAI techniques facilitate the identification of
attack vectors, aiding in the development of
more robust models. Through experimental
evaluations and case studies, we demonstrate
the efficacy of XAI in enhancing model
security and reliability. The findings
underscore the importance of integrating
Explainable AI strategies into the model
development pipeline to bolster defence
mechanisms against data poisoning attacks,
thereby fostering trust and dependability in
AI systems across diverse applications

A Data Quality Study Investigates How to Achieve Excellent Data Quality Inside an Organisation

Sanjay Kumar — Thu, 14 Dec 2023 00:00:00 +0530

The goal of this paper is to identify and
explore future difficulties in data quality
research opportunities for achieving high
data quality inside an organisation. The
systematic literature review method was used
for the review, which was based on research
papers released in conference proceedings
and journals. We devised an approach for
reviewing key issues such as current data
quality research fields, critical dimensions in
data quality, models and techniques for data
quality management, as well as ways for
assessing data quality. We select relevant
research publications based on the evaluation
method, collect and synthesize data, and
answer our research questions. The report
analyses the gaps in future research and
shows the progress in data quality research
towards real-world applications Management
of organisations, the impact of data quality on
companies, and other research fields
Database-related data quality technology
solutions dominated the early years of data
quality research. However, because the
Internet has replaced newspapers as the
primary source of information, new study
topics such as web data quality evaluation
and big data are unavoidable. In addition,
this analysis highlights and examines essential
data quality characteristics in organisations
Data completeness, consistency, accuracy, and
timeliness are some examples.

Countering Salami Attacks with Hybrid Machine Learning Models

Tutta Naga Venkata Durga, Geetha Usha Sri Bade, Manas Kumar Yogi — Sat, 16 Dec 2023 00:00:00 +0530

Salami attacks, a subtle form of malicious
activity wherein adversaries execute
incremental actions to achieve their goals
while remaining unnoticed, pose a significant
threat in various domains. This paper
introduces a novel approach to mitigate
salami attacks through the integration of
hybrid machine learning models. Leveraging
the strengths of both supervised and
unsupervised learning, our proposed
framework aims to enhance the detection and
prevention capabilities against these
clandestine attacks. The hybrid model
synergistically combines the interpretability
of supervised learning with the adaptability of
unsupervised learning, providing a
comprehensive solution for identifying
incremental threats across diverse datasets.
By analyzing patterns, anomalies, and
deviations from normal behaviour, the model
establishes a robust defence mechanism
capable of discerning subtle slices of
malicious activities characteristic of salami
attacks. Furthermore, our approach
incorporates real-time learning mechanisms
to adapt to evolving attack strategies,
ensuring continuous efficacy in countering
dynamic threats. Through extensive
experimentation on diverse datasets and
simulated attack scenarios, our hybrid
machine-learning model demonstrates
superior accuracy and efficiency compared to
traditional methods. This research
contributes to the ongoing efforts to fortify
cybersecurity measures, offering a proactive
and adaptable solution to safeguard against
the nuanced nature of salami attacks in
contemporary digital environments

Comparative Analysis of Various Online Fraud Detection Techniques and their Challenges

Deepti Ekka, Vinay Kumar Singh — Thu, 21 Dec 2023 00:00:00 +0530

Online fraudulent transactions are a
significant criminal violation. People and
financial institutions lose billions of dollars
every year. It highlights the importance of
financial institutions in detecting and
preventing fraud. Machine Learning
algorithms provide aprocess that can prevent
online business fraud with high accuracy.
Online fraud poses a significant threat to ecommerce and financial services. This study
explores the application of supervised
learning techniques for online fraud
detection. The primary objective is to build a
robust and accurate system that can
distinguish between legitimate and fraudulent
transactions in real-time. To achieve this goal,
we collect historical transaction data,
including relevant featuressuch as transaction
amount, location, time, and user behaviour.
This data is meticulously pre-processed to
handle missing values, outliers, and data
quality issues. Feature engineering is
performed to create new variables that
enhance the model's predictive power.
The dataset is divided into a training set; test
set and test set to facilitate the design process.
Various supervised learning algorithms
included logistic regression, decision tree;
random forest and support vector machines
consider finding a lie. The selected model is
trained on the labelled examples in the
training set, learning to discriminate between
fraudulent and legitimate transactions.
Perform hyperparameter tuning to optimize
model performance and Chapter F1 results.
When the model's performance meets
predefined criteria, it is sent to: The online
system allows it to record new changes in realtime.