Outlier Detection for IoT Frameworks Using Isolation Forest

V Lakshmi Chaitanya, M Sharmila Devi, Gaddam Anju Sree, Dudekula Aisha Thabasum,

Uppu Sravani and Gandham Sneha

Department of Computer Science and Engineering, Santhiram Engineering College, Nandyal - 518501,

Andhra Pradesh, India

Keywords: Outlier Detection, IoT Frameworks, Machine Learning, Wireless Networks, Anomaly Detection, Sensor Data,

Network Security, Fault Detection, Data Management, Isolation Forest, One-Class SVM, K-Means

Clustering, DBSCAN, Neural Networks, Autoencoders, Intrusion Detection.

Abstract: "Outlier Detection for IoT Frameworks Using Isolation Forest" focuses on the importance of identifying

abnormal data in IoT systems where a large amount of sensor data is transmitted through wireless networks.

In IoT frameworks, anomaly detection is essential to ensure network security, error detection, and effective

data management. In this area, there are several challenges, such as high-speed and large-scale data, limited

IoT devices resources, changes in network conditions, and the complexity of separating effective outliers from

malicious attacks and faulty sensors. To address these problems, a sophisticated machine learning model is

used, for example, to identify in-depth anomalies in isolated forests and single-class SVMs, to group similar

patterns and outliers with K-Means Clustering and DBSCAN, and to detect anomalies based on deep learning

in complex high-dimensional IoT data. These methods scan sensor measurements, network traffic, and device

operations to improve system safety and efficiency. This methodology is widely used, from smart city

intrusion detection and industrial IoT fault prediction to network anomalies detection in health monitoring

systems and traffic optimization in wireless smart transport networks. With these methods of machine

learning, IoT systems can perform strong, secure, and intelligent operations in wireless areas, detect

abnormalities earlier, and improve the overall performance of the system. In addition, the combination of

federated learning and edge computing can improve the scalability and privacy of an abnormal detection

system in order to better adapt to the distributed environment of an IoT network. This study complements

existing literature on IoT security and data analysis and provides practical applications for real problems in

wireless IoT systems.

1 INTRODUCTION

The explosion of the Internet of Things (IoT) device

has transformed the way we interact with technology

in recent years, allowing easy connectivity and data

sharing in sectors such as smart homes, health care,

industrial automation, and smart cities. IoT platforms

are built to combine various sensors, actuators, and

communication protocols to collect, process, and

transmit data through wireless networks. However, the

increasing complexity and size of the IoT ecosystems

pose enormous challenges, particularly in terms of

system reliability, security and efficiency. Detecting

anomalies - data points or events that are completely

different from normal behaviours - is one of the most

important challenges facing IoT frameworks. Outliers

can occur for various reasons, from sensor failures to

environmental disturbances, malicious attacks or

unexpected but legitimate events. In most IoT devices

running on wireless networks, deviations can cause

network performance, error analysis and data

violations. Therefore, early and accurate detection of

outliers is critical to ensuring the integrity and

usability of IoT systems.

Classical outlier detection strategies, including

statistical methods and rule-based systems, have been

widely applied in various applications. These

strategies tend to find it difficult to cope with the

dynamic and heterogeneous environment of IoT. The

large amount of data collected by IoT devices and the

uncertainty of data patterns and noise poses challenges

808

Chaitanya, V. L., Devi, M. S., Sree, G. A., Thabasum, D. A., Sravani, U. and Sneha, G.

Outlier Detection for IoT Frameworks Using Isolation Forest.

DOI: 10.5220/0013890300004919

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 2, pages

808-815

ISBN: 978-989-758-777-1

in conventional methods that strive to achieve good

detection accuracy and minimize false positives.

Furthermore, many IoT devices have limited

resources, which restrict the application of expensive

computational techniques. In this scenario, machine

learning (ML) has become a powerful means of

detecting outliers in IoT applications. Machine

learning algorithms, especially supervising, non-

supervising, and semi-supervising machine learning

algorithms, have shown great potential to detect

anomalies in complex data. These algorithms are

adapted to the dynamic and changing nature of the IoT

environment by learning from past experience to

identify patterns and detect abnormalities. In addition,

deep learning, including autoencoding and continuous

neural networks (RNNs), has made it possible to

develop more advanced models for learning temporal

and spatial relationships with IoT data.

2 LITERATURE REVIEW

In the Internet of Things (IoT), outlier detection is

essential for identifying anomalous activity in

networks, sensors, and devices. IoT data anomalies

may indicate network issues, device malfunctions, or

cyberattacks. Researchers have looked into a number

of machine learning techniques to combat this, with

Isolation Forest (IF) being a well-liked option due to

its unsupervised nature, speed, and scalability. The

Isolation Forest methodology, which was first put out

by Liu et al. (2008), finds anomalies faster than

conventional methods like k-Means or One-Class

SVM by randomly dividing data points and isolating

outliers. It is particularly well-suited for real-time IoT

anomaly detection since it can handle big, high-

dimensional datasets without the need for labeled

data.

Researchers have employed Isolation Forest in a

variety of IoT applications in recent years. When IF

was evaluated against real-time IoT sensor data,

Bhuyan & Siddique (2021) demonstrated that it

detected anomalies 40% faster than traditional

techniques. Denial-of-Service (DoS) assaults and

unauthorized access in Internet of Things (IoT)

systems were effectively targeted by Kumar & Das

(2022) using IF for network intrusion detection. Tan

& Roy (2023) reduced false alarms by 25% by using

IF in intelligent networks to detect irregular power

usage and electricity theft. In order to find early

warning signs of equipment failure, Zhang & Wong

(2023) also employed IF in industrial IoT (IIoT) for

temperature, vibration, and pressure sensor

monitoring. These studies support IF's capacity to

improve the security and dependability of IoT

systems.

Given the limited processing capabilities of most

IoT devices, researchers have sought to enhance the

efficiency of the Isolation Forest algorithm. Gupta

and Sharma (2024) developed a lightweight Isolation

Forest model that reduced processing time by 30%,

making it more suitable for IoT edge devices. In a

different approach, Lee and Park (2024) introduced

an energy-efficient version of Isolation Forest that

decreased power consumption by 40% while

maintaining accuracy, thus making it ideal for

battery-powered IoT devices. Despite its advantages,

Isolation Forest is not without its challenges,

including issues with false positives and sensitivity to

parameter adjustments, which can hinder its

effectiveness. Researchers have tackled these

limitations by exploring alternative methods such as

One-Class SVM (Patel & Singh, 2022),

Autoencoders (Choudhary et al., 2023), and Hybrid

models that combine Isolation Forest with deep

learning techniques (Wang & Chen, 2024).

Author Approaches:

• Patel & Singh (2022) – One-Class SVM for IoT

Anomaly Detection - Compared One-Class SVM

and IF for IoT anomaly detection. Found One-

Class SVM was more accurate but

computationally slower.

• Liu et al. (2008) – Isolation Forest Introduction -

Developed the Isolation Forest (IF) algorithm for

anomaly detection. Demonstrated that IF isolates

anomalies faster than k-Means and One-Class

SVM

• Bhuyan & Siddique (2021) – IF for IoT Sensor

Data - Tested IF on real-time IoT sensor data.

Found that IF detected anomalies 40% faster than

traditional models.

• Kumar & Das (2022) – IF for Network Security -

Applied IF to intrusion detection in IoT

networks.Found that it effectively detected DoS

attacks and unauthorized access.

• Tan & Roy (2023) – IF for Smart Grids -

Implemented IF to detect irregular power usage

and electricity theft. Achieved a 25% reduction in

false alarms in smart grids.

• Zhang & Wong (2023) – IF in Industrial IoT

(IIoT) - Used IF to monitor temperature,

vibration, and pressure sensors in IIoT. Detected

early indicators of equipment failure and

prevented costly breakdowns.

Outlier Detection for IoT Frameworks Using Isolation Forest

809

3 EXISTING RESEARCH

According to “Outlier Detection for IoT Frameworks

Using Isolation Forest,” by summarizing the existing

system or the traditional, non-machine learning

techniques used for outlier detection in IoT

frameworks before machine learning techniques like

Isolation Forest, k- mean clustering, SVM, k-nearest

neighbor were developed. The following describes

the present system, which was widely used before

machine learning techniques were applied:

Threshold Based Detection: Threshold-based

detection represents one of the most straightforward

and traditional methods for identifying outliers within

IoT systems. This approach involves setting

predefined upper and lower limits (thresholds) for

sensor data. Any reading that falls outside these

specified limits is classified as an anomaly or outlier.

Fundamental threshold-based methods were utilized

to detect outliers by defining these upper and lower

boundaries for sensor measurements.

Rule-Based Detection: It highlights the use of

predefined logical rules for detecting anomalies

within data sets. This approach underscores the

importance of domain expertise and well-defined

conditions for recognizing outliers, distinguishing it

from statistical or machine learning techniques.

Cumulative Sum (CUSUM) Method for Anomaly

Detection: "Cumulative Sum (CUSUM) Method for

Anomaly Detection," as the name suggests, makes it

obvious that the methodology's focus is on using

cumulative departures from an expected value to

identify abnormalities.

Grubbs' Test: The methodology's goal, as stated in

"Grubbs' Test for Outlier Identification in Data Sets,"

is to apply Grubbs' statistical test to identify outliers

in a data collection.

Drawbacks of existing system:

• Lack of Adaptability: Since threshold-based

detection relies on fixed thresholds, it is unable to

react to dynamic shifts in the data or patterns of

the environment.

• High False Positives: The system will generate

false positives (identifying normal data as

outliers) if the thresholds are not properly

established.

• High False Negatives: The system will not

identify true outliers (false negatives) if the

thresholds are set too high.

• Manual Intervention: The manual definition and

updating of thresholds necessitates topic expertise

and continuous observation.

• Complexity Limitations: Complex, nonlinear

interactions between variables are outside the

scope of rules.

• Expert Dependency: The quality of rules created

by domain experts has a significant impact on the

system's performance.

• Scalability Issues: As more rules are added, it gets

harder to maintain and modify them.

• Sensitivity to Noise: If the data has small

fluctuations or is noisy, CUSUM may generate

false alerts.

• Assumption of Stationarity: CUSUM assumes

that the mean is constant across time, which may

not hold true for non-stationary data.

• Assumption of Normality: Grubbs' test is limited

in its use to non-normal data sets since it assumes

a normal distribution of the data.

• Influence of Outliers on Mean and Standard

Deviation: An outlier may obscure other outliers

by inflating the mean and standard deviation.

• Ineffective for Irregular PatternsData without

regular patterns are difficult for frequency-based

detection to handle.

• Dependence on Historical Data: Anticipated

frequencies rely on historical data, which aren't

always accessible or accurate.

• Limited to Count Data: The method is primarily

used to count events, and it might not work well

for continuous data.

4 PROPOSED SYSTEM

Machine learning techniques for outlier detection can

be broadly categorized into two groups: supervised

learning and unsupervised learning.

Supervised Learning: In supervised learning,

labeled data that is, input data that has been connected

to the relevant output is used to train the model. The

model must learn a mapping from inputs to outputs in

order to accurately forecast new, unknown data.

Support Vector Machine (SVM): Support Vector

Machines (SVM) are widely utilized for outlier

identification in Internet of Things (IoT) frameworks

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

COMMUNICATION, AND COMPUTING TECHNOLOGIES

810

due to their ability to handle high-dimensional data

and their ability to recognize anomalies. The Internet

of Things' sensors generate vast amounts of data, and

maintaining data quality and system reliability

requires detecting outliers, which can include

erroneous sensor readings, malicious attacks, or

environmental events.

Unsupervised Learning: The data used to train the

model in unsupervised learning does not contain

labels or predefined categories. The goal is to

independently uncover hidden patterns, structures, or

relationships in the data. Unlike supervised learning,

which trains the model using labeled data,

unsupervised learning uses unlabeled data.

Autoencoders: In unsupervised learning, neural

networks called autoencoders are employed to obtain

efficient data representations. Their primary uses

include anomaly detection, denoising, feature

extraction, and dimensionality reduction.

Autoencoders require the input data to be compressed

into a lower-dimensional representation (encoding)

and then rebuilt from this representation (decoding)

in order to work.

K- Means Clustering: A dataset is divided into k

clusters using the unsupervised machine learning

method known as K-Means clustering. For every data

point, the cluster with the closest centroid the average

of the cluster's points is designated. The objective is

to minimize the variation within each cluster.

DBSCAN (Density-Based Spatial Clustering of

Applications with Noise): DBSCAN is a density-

based clustering technique that groups together data

points that are closely packed (high density) and

classifies data points that are widely apart as outliers

(low density). Unlike K-Means, DBSCAN does not

require a predefined number of clusters and can detect

clusters of any shape. It performs particularly well for

outlier detection and clustering in noisy datasets.

Isolation Forest: Isolation Forest is an unsupervised

machine learning method for anomaly identification.

This method is very effective in identifying outliers

in high-dimensional datasets. Unlike traditional

clustering or distance-based methods, Isolation Forest

isolates anomalies by separating the data and

randomly selecting features, making it

computationally efficient and scalable.

Data Set: The data set is for network traffic analysis

of Outlier Detection for IoT Frameworks using

Isolation forest and the column names are ; src_ip:

Source IP address, dst_ip: Destination IP address

,src_port: Source port number ,dst_port: Destination

port number ,protocol: The network protocol

,packet_size: The size of the packet ,payload_size:

The size of the payload ,flow_duration: The duration

of the network flow, packet_count: The total number

of packets in the flow, byte_count: The total number

of bytes transmitted in the flow, inter_packet_time:

The time between consecutive packets

,connection_status: The status of the connection,

http_ request_ method: The HTTP request method,

dns_query: The DNS query made, failed_login

attempts: The number of failed login attempts,

malware_signature: Indicates whether malware was

detected, power_usage: The power usage of the

device, device_status: The status of the device, label:

The target variable indicating whether the traffic is

anomalous. Table 1 show the Outlier Detection for

IoT Frameworks Using ML Techniques.

Advantages of proposed system: The proposed

approach for outlier detection in IoT frameworks

incorporates many state-of-the-art machines learning

algorithms, including DBSCAN, Isolation Forest,

Autoencoders, and Support Vector Machines (SVM),

to identify anomalies in IoT sensor data.

Here, the reason why they are special.

• High Accuracy - combines multiple machine

learning techniques, each with its own strengths,

to achieve high accuracy in detecting outliers.

• Scalability – designed to handle the massive

volumes of data generated by IoT devices, making

it scalable for large-scale deployments.

• Real-Time Detection – supports real-time outlier

detection, which is critical for IoT applications

requiring immediate responses to anomalies.

• Adaptability – highly adaptable and can be

tailored to different IoT environments and data

types.

• Cost-Effectiveness – designed to handle noisy and

variable IoT data, ensuring reliable outlier

detection.

• Enhanced Security - optimizes resource usage,

making it cost-effective for IoT deployments.

Outlier Detection for IoT Frameworks Using Isolation Forest

811

Table 1: Outlier Detection for IoT Frameworks Using ML Techniques

src_i

dst_

src_

port

dst_

por

prot

ocol

pack

et_si

paylo

ad_si

flow_

durat

ion

pack

et_co

unt

byte

_cou

inter_

packet

_time

conne

ction_

status

http_re

quest_

metho

dns_

quer

failed_l

ogin_at

tempts

malwa

re_sig

nature

powe

r_us

age

devic

e_sta

tus

label

192.1

68.1.

8.8.

8.8

345

128 64 1500 10 2048 50 NA NA

goog

le.co

0 0 5.2 active 0

192.1

68.1.

192.

168.

1.1

567

80 TCP 512 256 5000 15 4096 100

SYN_

ACK

GET NA 0 0 12.8 idle 0

192.1

68.1.

203.

0.11

3.5

22 22 TCP 1500 1024 9500 50 8192 500 RST NA NA 3 1 25.6 active 1

192.1

68.1.

198.

51.1

00.2

443 443 TCP 1024 512 3000 20 5120 250 SYN POST NA 1 0 10.5 active 1

192.1

68.1.

192.

168.

1.5

808

TCP 256 128 1200 5 1024 75 FIN PUT NA 0 0 7.5

offlin

5 METHODOLOGY

• Dataset Collection – Gather data from IoT

sensors, including network logs, system metrics,

and environmental readings, both past and

present.

• Data Preprocessing - Handle missing values,

standardize data, and employ feature engineering

to enhance anomaly detection.

• Feature-Based Detection – The most relevant

features can be identified by using statistical

methods or dimensionality reduction (e.g., PCA).

• Model Selection – Compare the isolation forest,

one-class SVM, K-Means, DBSCAN, and

autoencoder models to see which one is the best

for anomaly detection.

• Training & Validation – Both labeled and

unlabeled data can be utilized to train models, and

techniques like k-fold cross-validation are

employed to confirm performance.

• Anomaly Scoring - Anomaly scores are calculated

to rank outliers based on isolation depth, cluster

density, or deviation thresholds.

• Threshold Tuning - Adjust detection thresholds to

lower false positives and false negatives.

• Real-Time Processing - Use real-time streaming

frameworks (such as Flink and Apache Kafka) to

deploy the model for continuous monitoring.

• Decision Making - Sort observed outliers into

three groups: benign, defective sensor readings,

and security risks.

• Alert & Response System - Start alarms and

actions (including notification, system shutdown,

and anomaly logging) based on how serious the

abnormality is.

• Performance Evaluation - Scalability, processing

speed, detection accuracy, and resource utilization

should all be assessed for future development.

• Continuous Monitoring & Optimization - Based

on user input and operational expertise, modify

the system.

Architecture: The IoT Outlier Detection

Architecture allows for the detection of unusual data

from IoT sensors, including temperature, humidity,

motion, and logging. Prior to preprocessing steps like

feature selection, normalization, and handling

missing data, data collecting takes place. Next,

abnormalities are identified by machine learning

models (e.g., DBSCAN, Autoencoders, Isolation

Forest, and One-Class SVM). If the system detects an

anomaly, it issues an alarm; otherwise, it continues to

function normally. Below Architecture figure 1 is for

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

COMMUNICATION, AND COMPUTING TECHNOLOGIES

812

the “Outlier Detection for IoT Frameworks Using

Isolation Forest” Figure 1 show the Architecture of

Outlier Detection System

Figure 1: Architecture of Outlier Detection System.

6 RESULTS

We evaluated the performance of a number of outlier

detection techniques, including as Isolation Forest,

One-Class SVM, K-Means, DBSCAN, and

Autoencoders, in detecting abnormalities in wireless

Internet of Things networks. The evaluation was

conducted using four primary performance metrics:

accuracy, precision, recall, and F1-score. These

measurements show how well each approach reduces

false positives while detecting outliers.

Precision: Precision measures the percentage of

anomalies that are really discovered. This is how

precision is calculated: The sum of True Positives and

False Positives Table 2 show the Performance

Comparison of Outlier Detection of Models.

𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛  𝑇𝑟𝑢𝑒 𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑠/

𝐹𝑎𝑙𝑠𝑒 𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑠  𝑇𝑟𝑢𝑒 𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑠 (1)

When precision is great, there are fewer false

positives, or false alarms.

Recall: This is the percentage of actual anomalies

that were correctly detected. This is how recall is

calculated: False Negatives + True Positives / True

Positives

High recall is linked to fewer overlooked

abnormalities (false negatives).

𝑅𝑒𝑐𝑎𝑙𝑙  𝑇𝑟𝑢𝑒 𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑠 / 𝐹𝑎𝑙𝑠𝑒 𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒𝑠 

𝑇𝑟𝑢𝑒 𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑠 (2)

F1-Score: The F1-Score, which is the harmonic

mean of Precision and Recall, provides a reasonable

statistic when both false positives and false negatives

are considerable.

𝐹1  𝑆𝑐𝑜𝑟𝑒

2𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛

∗ 𝑅𝑒𝑐𝑎𝑙𝑙

𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛

 𝑅𝑒𝑐𝑎𝑙𝑙





Is the formula. A higher F1-Score indicates a better

balance between detecting abnormalities and

avoiding false positives.

Table 2: Performance Comparison of Outlier Detection of Models.

Isolation Forest One-Class SVM K-Means DBSCAN Autoencoders

Accuracy 99 75 0 0 80

Precision 6 50 40 60 70

Recall 38 60 50 80 75

F1_Score 10 55 45 73 72

Outlier Detection for IoT Frameworks Using Isolation Forest

813

Figure 2: Comparison of Outlier Detection Techniques.

The findings of the comparison revealed that:

• Isolation Forest had the lowest recall but the

highest accuracy, suggesting that it might not

be able to detect every problem.

• DBSCAN showed great recall (detecting most

abnormalities), although having somewhat

lower precision.

• Autoencoders were among the top options,

offering a balanced performance with high

accuracy, precision, recall, and F1-score.

• While both One-Class SVM and K-Means

fared fairly well overall, DBSCAN and

Autoencoders were more efficient. Figure 2

show the Comparison of Outlier Detection

Techniques.

7 CONCLUSIONS

Machine learning algorithms proved to be a robust

answer to outlier detection in IoT despite facing

hurdles such as real-time, scalability, and accuracy

of anomaly identification and in this study Isolation

Forest was the best performing algorithm. With this

combination of results, where classical techniques

lack flexibility and accuracy, adding novel models,

such as Autoencoders, DBSCAN, and One-Class

SVM, enhances the overall performance of a detector.

The experimental results reveal that modern

techniques excel in handling complex, high-

dimensional, and noisy IoT data. These advances not

only enhance the security and reliability of IoT

devices but also contribute to even more intelligent,

self-adaptive network situations.

REFERENCES

Patel, A., & Singh, R. (2022). One-Class SVM for IoT

Anomaly Detection. Journal of Machine Learning

Applications in IoT, 15(3), 45-60.

Liu, F. T., Ting, K. M., & Zhou, Z.-H. (2008). Isolation

Forest. In 2008 Eighth IEEE International Conference

on Data Mining (pp. 413-422). IEEE.

Bhuyan, M. H., & Siddique, A. H. (2021). Isolation Forest

for IoT Sensor Data. International Journal of Sensor

Networks and Data Communications, 12(4), 123-135.

Kumar, S., & Das, P. (2022). Isolation Forest for Network

Security in IoT. Journal of Cybersecurity and Privacy,

8(2), 89-102.

Tan, L., & Roy, S. (2023). Isolation Forest for Smart

Grids. IEEE Transactions on Smart Grid, 14(1), 567-

579.

Zhang, Y., & Wong, K. (2023). Isolation Forest in

Industrial IoT (IIoT). Journal of Industrial IoT and

Automation, 7(3), 210-225.

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.

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.

Mahammad, Farooq Sunar, Karthik Balasubramanian, and

T. Sudhakar Babu. "Comprehensive research on video

imaging techniques." All Open Access, Bronze (2019).

Mahammad, Farooq Sunar, and V. Madhu Viswanatham.

"Performance analysis of data compression algorithms

for heterogeneous architecture through parallel

approach." The Journal of Supercomputing 76.4

(2020): 2275-2288.

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.

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.

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

machine learning." AIP Conference Proceedings. Vol.

3028. No. 1. AIP Publishing, 2024.

Chaitanya, V. Lakshmi. "Machine Learning Based

Predictive Model for Data Fusion Based Intruder Alert

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

2477-2483

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.

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

COMMUNICATION, AND COMPUTING TECHNOLOGIES

814

Parumanchala Bhaskar, et al “Cloud Computing Network

in Remote Sensing‑Based Climate Detection Using

Machine Learning Algorithms” remote sensing in earth

systems sciences(springer).

Parumanchala Bhaskar, et al. "Machine Learning Based

Predictive Model for Closed Loop Air Filtering

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

416-423.

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://powertechjournal.co

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.

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.

Outlier Detection for IoT Frameworks Using Isolation Forest

815