Smart Detection and Prevention of Cloud Based Security Threats

Using Machine Learning

Parumanchala Bhaskar, Somisetty Akhil, R. S. Venkhatesh, Shaik Sajid,

M. Prem Kumar and S. Mansoor Basha

Department of Computer Science and Engineering, Santhiram Engineering College, Nandyal, Andhra Pradesh, India

Keywords: Privilege Escalation, Insider Attack, Machine Learning, Random Forest, Adaboost, XGBoost, LightGBM,

Classification.

Abstract: This task utilizes advanced machine learning to enhance cloud protection, specifically addressing and

mitigating privilege escalation attacks for a greater resilient protection mechanism. With the growth in cloud

adoption, the threat of privilege escalation attacks also escalates. This assignment goals holes in employee get

admission to privileges in cloud services to improve usual protection. The task makes use of machine learning

to facilitate actual-time identification and prevention of privilege escalation attacks. Strategies like as

LightGBM, Random forest, Adaboost, and Xgboost provide a sturdy protection in opposition to rising threats.

Individuals and establishments come upon stronger records protection, cultivating self-assurance in cloud

computing. Cloud service vendors and corporations accumulate guarantee in a comfortable online ecosystem,

reaping the advantages of the task's security upgrades. A vote casting Classifier, integrating predictions from

decision Tree, Random forest, and support Vector machine through a "tender" vote casting method, improves

the system's efficacy in figuring out and countering privilege escalation attacks. A person-friendly Flask

framework with SQLite integration enhances user trying out by way of offering secure signup and signin

functionalities for realistic installation and evaluation.

1 INTRODUCTION

Cloud computing represents a progressive paradigm

for delivering and facilitating offerings through the

internet. The present infrastructure. Cloud storage

organizations put into effect essential security

protocols for their structures and the facts they

manage, encompassing encryption, get entry to

control, and authentication. The cloud gives nearly

infinite ability for storing numerous varieties of data

throughout diverse data storage architectures,

contingent upon accessibility, velocity, and

frequency of statistics retrieval. Data breaches

involving sensitive facts may additionally get up from

the big flow of facts between organisations and cloud

carrier vendors, whether unintended or intentional.

Parumanchala Bhaskar, et al., The attributes that

facilitate consumer-friendliness in online services for

employees and IT structures simultaneously

complicate efforts for companies for preventing

unauthorized access. The Authentication and open

interfaces constitute rising protection issues that

agencies come across with cloud services. Notably

professional hackers employ their information to

infiltrate Cloud infrastructure. Parumanchala

Bhaskar, et al. 2022; Mahammad, Farooq Sunar, et al.

2024, Machine learning to know utilizes numerous

methodologies and algorithms to address security

challenges and decorate data control. Numerous

datasets are confidential and can't be disclosed due to

privacy issues, or they will lack essential statistical

characteristics.

The rapid expansion of the Cloud area engenders

privacy and protection threats regulated via rules.

Worker access privileges may additionally remain

unchanged as they transition to different jobs or

positions inside the Cloud agency. Consequently,

obsolete privileges are exploited detrimentally to

suitable and harm precious data. Every account that

interacts with a computer possesses a positive

diploma of authority. Server databases, sensitive

statistics, and additional services are regularly

restrained to authorized customers. An adverse

attacker can infiltrate a touchy system by seizing

696

Bhaskar, P., Akhil, S., Venkhatesh, R. S., Sajid, S., Kumar, M. P. and Basha, S. M.

Smart Detection and Prevention of Cloud Based Security Threats Using Machine Learning.

DOI: 10.5220/0013871500004919

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 1, pages

696-704

ISBN: 978-989-758-777-1

manipulate of a privileged user account and

leveraging or augmenting privileges. According to

Parumanchala Bhaskar, et al, 2024 their desires,

attackers may additionally maneuver horizontally to

advantage manage over extra structures or vertically

to reap administrative and root get entry to until they

achieve complete manipulate over the complete

surroundings. Horizontal privilege escalation occurs

while a person acquires the access permissions of

another person owning the equal get entry to stage. A

culprit may additionally appoint horizontal privilege

escalation to reap information that isn't always

inherently associated with them. An attacker can also

make the most vulnerabilities in a poorly built web

application to get entry to the personal data of others.

The attacker has successfully completed a horizontal

privilege escalation make the most, permitting them

to view, adjust, and duplicate sensitive records.

Adversaries focus on facts repositories due to

their possession of the maximum valuable and touchy

facts. The privacy and security of each cloud user are

compromised if data is misplaced. Insider threats are

unfavourable activities performed by people with

authorized get admission to. Because of the rapid

growth of networks, numerous corporations and

groups have advanced own internal networks. Latest

estimates imply that 90% of corporations perceive

themselves as prone to insider threats. Malefactors

can make the most privilege escalation to create

supplementary assault vectors on a target system.

Insider attackers searching for to attain accelerated

privileges or get right of entry to more touchy systems

via privilege escalation tries. Insider attacks are

difficult to stumble on and thwart because of their

operation below company-level safety protocols and

often possess privileged get entry to the network.

Identifying and categorizing insider threats has come

to be hard and labor-in depth.

Latest research focused at the detection and class

of privileged elevation attacks perpetrated by insider

people. They advised numerous machines getting to

know and deep learning methodologies to address

those problems. Latest studies employed strategies

consisting of "SVM, Naïve Bayes, CNN, Linear

Regression, PCA, Random forest, and KNN". The

requirement for speedy and efficient machine

learning algorithms is greatly esteemed because to the

form of assault types. Therefore, a robust and

efficient approach is vital to discover, categorize, and

alleviate those insider threats. mareswara Kumar,

2024, To enhance protection protection systems, it is

essential to hire wise algorithms, which include

machine learning algorithms, for the class and

prediction of insider attacks.

Moreover, know-how the efficacy of machine

learning algorithms in categorizing insider assaults

enables the choice of the most appropriate algorithm

for every state of affairs, necessitating enhancements

to the algorithms themselves. This permits the supply

of more suitable security features. This study seeks

to enforce powerful and efficient machine learning

algorithms in insider attack conditions to get

advanced and expedited consequences. Machine

learning methods were implemented and assessed in

this context: "Random forest area, AdaBoost,

XGBoost, and LightGBM". The boosting strategy's

approach involves improving a weak classifier to

seriously improve its performance through

augmenting the classification set of rules's

predictions. "Random forest, AdaBoost, and

XGBoost" efficaciously and correctly categorized

insider threats.

2 LITERATURE REVIEW

Cloud computing refers to the on-call for accessibility

of computing infrastructure resources. Particularly

the capacity for information storage and management,

without direct, exclusive oversight by means of the

user. It has offered clients each public and personal

computing and data garage on a unified platform via

the net. In addition, it encounters severa safety risks

and challenges which can impede the adoption of

cloud computing answers. Suman, Jami Venkata, et

al. 2024 This paper discusses the safety troubles,

challenges, methods, and solutions associated with

cloud computing. a multitude of individuals

expressed safety apprehensions in a prior poll. Every

other survey examines the cloud computing

architecture approach, with several detailing

protection worries and methodologies. This text

consolidates all safety issues, challenges,

methodologies, and answers in a single vicinity.

Cloud computing denotes the instantaneous

accessibility of private computer system sources,

specifically statistics storage and processing

capabilities, impartial of the client's intervention.

Emails are often applied for the transmission and

reception of statistics amongst individuals or

businesses. Financial facts, credit reviews, and other

touchy facts are often transmitted over the internet.

Phishing is a fraudulent method employed to obtain

touchy statistics from people via masquerading as

professional sources. The sender can manipulate you

into divulging exclusive data via deception in a

phishing email. The primary difficulty is email

phishing assaults at some stage in the transmission

Smart Detection and Prevention of Cloud Based Security Threats Using Machine Learning

697

and reception of emails. The assailant transmits spam

content by means of e mail and acquires your facts

upon your commencing and reading of the email. In

current years, it has posed a good-sized undertaking

for all. This study employs varying quantities of

legitimate and phishing datasets, identifies clean

emails, and utilizes numerous attributes and

algorithms for categorization. A revised dataset is

generated next to comparing the contemporary

methodologies. Mahammad, et al. 2020; Sharmila, et

al. 2022, We generated a characteristic-extracted

"comma-separated values (CSV)" record and a label

report, then implemented the "support vector machine

(SVM), Parumanchala Bhaskar, et al. 2024; D. C. Le,

et al., 2020, Naive Bayes (NB), and long short-term

memory (LSTM)" algorithms. This test regards the

identification of a phished email as a category trouble.

The comparison and implementation imply that

"SVM, NB, and LSTM "exhibit advanced

performance and accuracy in detecting e-mail

phishing assaults. E-mail attack class using "SVM,

NB, and LSTM" classifiers attained most accuracies

of ninety-nine. Sixty-two%, 97%, and ninety-eight%,

respectively.

With trends in technological know-how and

generation, cloud computing represents the

approaching huge development inside the industry.

Cloud cryptography is a way that use encryption

techniques to guard information. The primary benefit

of cloud storage is its accessibility, decreased

hardware requirements, decrease upkeep, and

protection prices, main to widespread adoption

through corporations. Encryption is the technique of

encoding information to inhibit unauthorized get

admission to. Currently, we aim to guard the data

stored on our computer systems or transferred

through the net from assaults. 4 The cryptography

method relies on response pace, confidentiality,

bandwidth, and integrity. Moreover, safety is a

critical factor of cloud computing for ensuring the

safety of client facts on the cloud. Our research

article evaluates the efficiency, application, and value

of existing cryptographic methods. The evaluation

outcomes imply the most suitable method for

particular records kinds and environments.

The extensive usage of era nowadays has given

rise to numerous security issues. Both public and

industrial sectors allocate a significant amount of

their budgets to protect the confidentiality, integrity,

and availability of their data towards ability attacks.

Insider assaults are more severe than external attacks,

as insiders are legal people with lawful get right of

entry to an agency's sensitive assets. Consequently,

numerous studies in the literature awareness on

developing techniques and gear to identify and

mitigate numerous forms of insider threats. This

newsletter evaluates several procedures and defenses

presented to thwart insider attacks. A complete type

version is proposed to categorize insider danger

prevention techniques into two types: biometric-

primarily based and asset-based metrics. The

biometric category is assessed into physiological,

behavioral, and physical sorts, at the same time as the

asset metric category is classified into host, network,

and aggregate kinds. This class organizes the tested

methodologies which are substantiated via empirical

findings via the grounded principle method for an

intensive literature take a look at. The item

additionally compares and analyzes essential

theoretical and empirical elements that substantially

have an effect on the efficacy of insider hazard

prevention techniques, which includes datasets,

characteristic domain names, type algorithms,

assessment metrics, real-world simulations, balance,

and scalability. Sizable barriers are emphasized that

need to be addressed while imposing real-

international insider hazard prevention systems.

Numerous studies gaps and proposals for destiny

research directions also are delineated.

The net of things is an advancing technology in

which interconnected computing devices and sensors

change records throughout a network to analyze

diverse troubles and offer novel services. The

"internet of things (IoT)" serves because the

fundamental permitting era for smart homes. Clever

home technology offers several capabilities to users,

along with temperature monitoring, smoke detection,

automatic lighting control, and smart locks.

However, it additionally introduces a new array of

security and privacy concerns; for instance,

unauthorized get entry to customers' personal facts

may additionally occur through the manipulation of

surveillance devices or the triggering of false

fireplace alarms, amongst other strategies. Sunar, et

al., 2018, Those challenges render clever houses

susceptible to several safety threats, inflicting

individuals to hesitate in adopting this technology due

to security issues. This survey record elucidates the

"internet of things (IoT)", its growth trajectory, item

standards, the layered architecture of the IoT

ecosystem, and the numerous security problems

associated with each layer in the smart domestic

context. This observe delineates the demanding

situations and issues arising in IoT-based smart

homes while also presenting techniques to mitigate

these security concerns.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

698

3 METHODOLOGY

3.1 Proposed Work

The advised system is a machine learning approach

for the identity and categorization of insider threats in

cloud settings. The utility of Random forest,

Adaboost, XGBoost, and LightGBM algorithms

improves predictive performance. The counseled

approach enhances accuracy in identifying insider

threats by way of using numerous machines learning

techniques, including "Random forest, Adaboost,

XGBoost, and LightGBM". The gadget employs

ensemble learning methods to combine the strengths

of many algorithms, therefore improving prediction

performance for insider danger detection in cloud

environments. The gadget makes use of

comprehensive data pretreatment methods, including

aggregation and normalization, to tackle issues which

include missing values, outliers, and extraneous

features, for this reason enhancing model

performance. Parameters including learning rate,

most depth, and k-fold are calibrated to enhance the

efficacy of machine learning models, facilitating an

extra powerful and customized approach for insider

threat identity. Figure 1. Show the Proposed

architecture A voting Classifier, integrating

predictions from decision Tree, Random forest, and

support Vector machine through a "soft" balloting

method, improves the system's efficacy in identifying

and countering privilege escalation assaults. A

consumer-pleasant Flask framework with SQLite

integration complements user testing by offering safe

signup and sign in functionalities for practical

installation and assessment

3.2 System Architecture

Figure 1: Proposed architecture.

The system architecture consists of four essential

tiers: facts collecting, data preprocessing,

implementation of supervised machine learning

algorithms, and consequences evaluation. During the

records gathering phase, a tailored dataset extracted

from various files of the CERT dataset is employed.

The collected data is subsequent subjected to

preprocessing, which includes methods like statistics

aggregation, normalization, and feature extraction to

improve its exceptional and relevance. The system's

foundation involves utilizing machine learning

algorithms "Random forest, AdaBoost, XGBoost,

and LightGBM" alongside a voting classifier as an

enhancement, to research the preprocessed statistics

for the identity and categorization of privilege

escalation assaults. The gadget does a comprehensive

evaluation of the results, assessing the efficacy of

each set of rules and imparting insights into the

complete system's success in detecting insider threats.

This structure guarantees a methodical and resilient

strategy for mitigating privilege escalation assaults

via device learning methodologies.

3.3 Dataset Collection

Chaitanya, V. et al 2022; 2014, The dataset utilized

on this experiment is sourced from several files

within the CERT dataset, with a specific emphasis on

e mail-related records.

Table 1. Show the CERT dataset

This curated dataset encompasses many cases

pertinent to insider threat scenarios in email

exchanges. It encompasses several characteristics

and properties referring to user conduct, email

content, and device interactions.

3.4 Data Processing

Data processing is converting unrefined data into

usable facts for corporations. Facts scientists

generally have interaction in facts processing,

encompassing the collection, corporation, cleaning,

validation, analysis, and transformation of records

into interpretable formats along with graphs or

papers. Facts processing may be carried out through

3 methods: manual, mechanical, and digital. The

objective is to beautify the value of facts and

streamline decision-making. This lets in enterprises

to decorate their operations and execute speedy

strategic decisions. Automated records processing

technologies, consisting of computer software

programming, are pivotal on this context. it could

remodel full-size datasets, specifically large facts,

into great insights for first-rate control and decision-

making.

Smart Detection and Prevention of Cloud Based Security Threats Using Machine Learning

699

Table 1: CERT Dataset.

date

user

bcc

{R317-

S4TX96FG-

8219JWFF}

01/02/2010

07:11:45

LAP

0338

PC-

5758

Dean.Flynn.Hines@dtaa.com;

Wade_Harrison@lockhe

Nathaniel.Hu

nter.Health@dt

aa.com

NaN

Lynn.Adena.

{R0R9-

E4GL59IK-

2907OSWJ}

01/02/2010

07:12:16

MOH

0273

PC-

6699

Odonell-Gage@belisouth.net

NaN

MOH68

{G2B2-

ABXY58CP-

2847ZJZL

}

01/02/2010

07:13:00

LAP

0338

PC-

5758

Penelope_colon@netzero.com

NaN

Lynn_A_Pra

{A3A9-

F4TH89AA-

8318GFGK

}

01/02/2010

07:13:17

LAP

0338

PC-

5758

Judith_Hayden@comcast.net

NaN

Lynn_A_Pra

{E8B7-

C8FZ88UF-

2946RUQQ}

01/02/2010

07:13:28

MOH

0273

PC-

6699

Bond-Raymond@verizon.net;

Alea_Ferrell@msn.com;

NaN

Odonnell-

gage@bell

MOH68

South.net

3.5 Feature Selection

Feature selection is the process of identifying the

most consistent, non-redundant, and pertinent

functions for model development. Systematically

minimizing dataset sizes is crucial even as the volume

and diversity of datasets persist in expanding. The

primary goal of feature selection is to enhance the

efficacy of a predictive model whilst minimizing the

computational rate of modeling.

Figure 2: LightGBM.

Feature selection, a critical aspect of function

engineering, involves identifying the most massive

features for enter into machine learning algorithms.

feature selection techniques are utilized to diminish

the quantity of input variables via excluding

redundant or pointless characteristics, hence refining

the function set to those most pertinent to the system

learning model. The number one blessing of

conducting feature selection ahead, instead of

allowing the machine learning version to decide the

most giant features autonomously.

3.6 Algorithms

LightGBM: "LightGBM" is a gradient boosting

ensemble method employed through the educate the

usage of AutoML tool, utilizing decision trees as its

basis. Figure 2. Show the LightGBM Much like other

decision tree-based techniques, "LightGBM" is

relevant for each category and regression tasks.

"LightGBM" is engineered for advanced overall

performance in distributed systems. XGBoost:

"XGBoost" operates as an efficient and widely-used

open-source implementation of the gradient boosted

trees technique within Amazon Sage Maker.

Figure 3.

Show the XGBoost

Gradient boosting is a supervised

learning technique that seeks to be expecting a target

variable accurately through aggregating the

predictions of a collection of smaller, weaker models.

Figure 3: XGBoost.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

700

AdaBoost: "AdaBoost, or Adaptive Boosting", is a

"machine learning" method employed as an

Ensemble method. Figure 4. Show the Adaboost the

predominant estimator hired in AdaBoost is a

decision tree with a single stage, indicating a decision

tree with only one break up. Those trees are known

as decision Stumps.

Figure 4: Adaboost.

RF: "Random forest" is an extensively applied

"machine learning" method, patented by Leo Breiman

and Adele Cutler that amalgamates the outputs of

numerous decision trees to get a singular outcome.

Figure 5. Show the Random forest Its person-

friendliness and adaptability have driven its

recognition, because it addresses each type and

regression issues.

Figure 5: Random forest.

VC: A voting Classifier is a "machine learning"

model that aggregates many models and predicts an

output class based on the best probability among them

Figure 6 show the Voting classifier.

Figure 6: Voting classifier.

4 EXPERIMENTAL RESULTS

Table 2: Performance evaluation.

Figure 7: Comparison graph.

Figure 7 So, this is the performance metrics

comparison graph.

The x-axis denotes algorithm names, while the y-

axis indicates performance measures.

Accuracy Recall Precision F1

LightGB

94.75 50 47.375

48.652

Xgboost 94.75 50 47.375

48.652

AdaBoos

95.45

58.016

90.27778

62.425

Random

Forest

95.45

58.016

90.27778

62.425

Voting

Classifier

96.45

66.640

96.82903

73.902

Smart Detection and Prevention of Cloud Based Security Threats Using Machine Learning

701

The blue bar signifies accuracy, the orange represents

recall, the inexperienced shows precision, and the red

corresponds to the F1 score.

Precision: Precision assesses the proportion of

accurately classified cases among the ones identified

as positive. Consequently, the formula for calculating

precision is expressed as:

“Precision = True positives/ (True positives + False

positives) = TP/ (TP + FP)” (1)

Precision 





(2)

Recall: recall is a metric in "machine learning' that

assesses a model's potential to recognize all pertinent

times of a specific class. It is the ratio of

appropriately anticipated fantastic observations to the

total real positives, offering insights right into a

version's efficacy in identifying occurrences of a

particular class.

Recall 





(3)

Accuracy: Accuracy is the ratio of correct

predictions in a classification test, assessing the

overall precision of a model's predictions.

Accuracy 





(4)

F1 Score: The F1 score is the harmonic suggest of

accuracy and recall, providing a balanced metric that

money owed for each false positive and false

negative, thereby making it appropriate for

imbalanced datasets.

F1 Score  2 ∗

  

  

* 100 (5)

Figure 8: Home page.

Figure 9: Signup page.

Figure 10: Signin page.

Figure 8 and 9 and 10 shows the home page signup page

and signin page respectively.

Figure 11: User input page.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

702

Figure 12: Prediction result.

Figure 11 and 12 show the User input page and

prediction result respectively.

5 CONCLUSIONS

The malevolent insider poses a critical threat to the

company because of their improved access and

chance to inflict substantial harm. Insiders have

exclusive and suitable access to knowledge and

resources, unlike outsiders. This look at present’s

machine learning algorithms for the detection and

classification of insider attacks. Chaitanya, et al.,

2014, This work utilizes a tailored dataset derived

from numerous documents of the CERT dataset. Four

machine learning strategies were implemented at the

dataset, yielding advanced results. The algorithms

consist of Random forest, AdaBoost, XGBoost, and

LightGBM. This research exhibited excellent

experimental findings with greater accuracy within

the categorization document the usage of supervised

machine learning methods. Of the proposed

algorithms, LightGBM achieves the maximum

accuracy at 97%, whilst Random Forest (RF) attains

86%, AdaBoost reaches 88%, and XGBoost records

88.27%. In the future, the proposed models might

also enhance their overall performance by means of

augmenting the dataset in both quantity and variety of

its residences and the emerging dispositions of insider

attackers. This can initiate novel research directions

for the detection and category of insider attacks

across many organizational domains. Groups make

use of machine learning models to facilitate informed

decision-making, and enhanced version

consequences bring about advanced judgments. The

price of errors might be enormous; but, this expense

diminishes with more suitable version precision.

machine learning-based totally research lets in

customers to supply sizable datasets to computer

algorithms, which sooner or later analyze, advise, and

make decisions based totally on the provided data.

6 FUTURE SCOPE

Future improvements need to prioritize increasing the

system's scalability to successfully control

heightened workloads in extensive cloud

environments, ensuring seamless processing as

records complexity and extent growth. Destiny

advancements must contain dynamic reaction

systems that can swiftly hit upon and counteract

newly developing strategies in privilege escalation

attacks, thereby imparting a proactive protection

against evolving insider threats. The incorporation of

methods that yield comprehensible justifications for

model decisions is important. R. Kumar., et al, 2020;

D. Tripathy., et al, 2020, This transparency aids

security analysts in understanding the factors

affecting threat identity, as a result enhancing trust

inside the system's consequences. It is critical to set

up a framework for the constant updating and

diversification of the dataset utilized for model

training. Continuous enhancement guarantees the

device's efficacy in detecting and countering novel

assault vectors and emerging insider threat trends.

REFERENCES

Ajmal, S. Ibrar, and R. Amin, ‘‘Cloud computing platform:

Performance analysis of prominent cryptographic

algorithms,’’ Concurrency Comput., Pract. Exper., vol.

34, no. 15, p. e6938, Jul. 2022.

Chaitanya, V. Lakshmi, and G. Vijaya Bhaskar. "Apriori vs

Genetic algorithms for Identifying Frequent Item Sets."

International journal ofInnovative Research

&Development 3.6 (2014): 249-254.

Chaitanya, V. Lakshmi. "Machine Learning Based

Predictive Model for Data Fusion Based Intruder Alert

System." journal of algebraic statistics 13.2 (2022):

2477-2483

Chaitanya, V. Lakshmi, et al. "Identification of traffic sign

boards and voice assistance system for driving.” AIP

Conference Proceedings. Vol. 3028. No. 1. AIP

Publishing, 2024

D. C. Le and A. N. Zincir-Heywood, ‘‘Machine learning

based insider threat modelling and detection,’’ in Proc.

IFIP/IEEE Symp. Integr. Netw. Service Manag. (IM),

Apr. 2019, pp. 1–6.

D. Tripathy, R. Gohil, and T. Halabi, ‘‘Detecting SQL

injection attacks in cloud SaaS using machine

learning,’’ in Proc. IEEE 6th Int. Conf. Big Data Secur.

Cloud (BigDataSecurity), Int. Conf. High Perform.

Smart Detection and Prevention of Cloud Based Security Threats Using Machine Learning

703

Smart Comput., (HPSC), IEEE Int. Conf. Intell. Data

Secur. (IDS), May 2020, pp. 145–150.

D. C. Le, N. Zincir-Heywood, and M. I. Heywood,

‘‘Analyzing data granularity levels for insider threat

detection using machine learning,’’ IEEE Trans. Netw.

Service Manag., vol. 17, no. 1, pp. 30–44, Mar. 2020.

Devi, M. Sharmila, et al. "Machine Learning Based

Classification and Clustering Analysis of Efficiency of

Exercise Against Covid-19 Infection." Journal of

Algebraic Statistics 13.3 (2022): 112-117.

Devi, M. Sharmila, et al. "Extracting and Analyzing

Features in Natural Language Processing for Deep

Learning with English Language." Journal of Research

Publication and Reviews 4.4 (2023): 497-502.

F. Janjua, A. Masood, H. Abbas, and I. Rashid, ‘‘Handling

insider threat through supervised machine learning

techniques,’’ Proc. Comput. Sci., vol. 177, pp. 64–71,

Jan. 2020.

G. Apruzzese, M. Colajanni, L. Ferretti, A. Guido, and M.

Marchetti, ‘‘On the effectiveness of machine and deep

learning for cyber security,’’ in Proc. 10th Int. Conf.

Cyber Conflict (CyCon), May 2018, pp. 371–390.

H. Touqeer, S. Zaman, R. Amin, M. Hussain, F. Al-

Turjman, and M. Bilal, ‘‘Smart home security:

Challenges, issues and solutions at different IoT

layers,’’ J. Supercomput., vol. 77, no. 12, pp. 14053–

14089, Dec. 2021.

Mahammad, Farooq Sunar, Karthik Balasubramanian, and

T. Sudhakar Babu. "A comprehensive research on

video imaging techniques." All Open Access, Bronze

(2019).

Mahammad, Farooq Sunar, and Y Madhu Viswanatham.

"Performance analysis of data compression algorithms

for heterogeneous architecture through parallel

approach." The Journal ofsupercomputing 76.4 (2020):

2275-2288.

Mahammad, Farooq Sunar, et al. "Key distribution scheme

for preventing key reinstallation attack in wireless

networks." AIP Conference Proceedings. Vol. 3028.

No. 1. AIP Publishing, 2024.

Mandalapu, Sharmila Devi, et al. "Rainfall prediction using

machine learning." AIP Conference Proceedings. Vol.

3028. No. 1. AIP Publishing, 2024.

Mr.M.Amareswara Kumar, “Baby care warning system

based on IoT and GSM to prevent leaving a child in a

parked car”in International Conference on Emerging

Trends in Electronics and Communication Engineering

- 2023, API Proceedings July-2024

Mr.M.Amareswara Kumar, effective feature engineering

technique for heart disease prediction with machine

learning” in International Journal of Engineering &

Science Research, Volume 14, Issue 2, April-2024 with

ISSN 2277-2685.

P. Oberoi, ‘‘Survey of various security attacks in clouds-

based environments,’’ Int. J. Adv. Res. Comput. Sci.,

vol. 8, no. 9, pp. 405–410, Sep. 2017.

Paradesi Subba Rao,” Detecting malicious Twitter bots

using machine learning” AIP Conf. Proc. 3028, 020073

(2024), https://doi.org/10.1063/5.0212693

Paradesi SubbaRao, "Morphed Image Detection using

Structural Similarity Index Measure”M6 Volume 48

Issue 4 (December 2024)

,https://powertechjournaI.com

Parumanchala Bhaskar, et al. "Machine Learning Based

Predictive Model for Closed Loop Air Filtering

System." Journal of Algebraic Statistics 13.3 (2022):

416-423.

Parumanchala Bhaskar, et al. "Incorporating Deep Learning

Techniques to Estimate the Damage of Cars During the

Accidents" AIP Conference Proceedings. Vol. 3028.

No. 1. AIP Publishing, 2024.

Parumanchala Bhaskar, et al “Cloud Computing Network

in Remote Sensing-Based Climate Detection Using

Machine Learning Algorithms” remote sensing in earth

systems sciences(springer).

R. Kumar, K. Sethi, N. Prajapati, R. R. Rout, and P. Bera,

‘‘Machine learning based malware detection in cloud

environment using clustering approach,’’ in Proc. 11th

Int. Conf. Comput., Commun. Netw. Technol.

(ICCCNT), Jul. 2020, pp. 1–7.

S. Zou, H. Sun, G. Xu, and R. Quan, ‘‘Ensemble strategy

for insider threat detection from user activity logs,’’

Comput., Mater. Continua, vol. 65, no. 2, pp. 1321–

1334, 2020.

Suman, Jami Venkata, et al. "Leveraging natural language

processing in conversational AI agents to improve

healthcare security." Conversational Artificial

Intelligence (2024): 699-711.

Sunar, Mahammad Farooq, and V. Madhu Viswanatham.

"A fast approach to encrypt and decrypt of video

streams for secure channel transmission." World

Review of Science, Technology and Sustainable

Development 14.1 (2018): 11-28.

U. A. Butt, R. Amin, H. Aldabbas, S. Mohan, B. Alouffi,

and A. Ahmadian, ‘‘Cloud-based email phishing attack

using machine and deep learning algorithm,’’ Complex

Intell. Syst., pp. 1–28, Jun. 2022.

U. A. Butt, R. Amin, M. Mehmood, H. Aldabbas, M. T.

Alharbi, and N. Albaqami, ‘‘Cloud security threats and

solutions: A survey,’’ Wireless Pers. Commun., vol.

128, no. 1, pp. 387–413, Jan. 2023.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

704