Third Year Project Finalization

THIRD YEAR PROJECT FINALIZATION REPORT

(2023- 2024)
This report, in a template, was designed for B.Tech. (CS) students working on their Third Year
Project at CS Department, ABES Engineering College, Ghaziabad.

Every group of students is required to do the following:

1. Complete all the sections of this template
2. Get it certified by the assigned internal advisor before the day of evaluation.
3. Submit original copy to coordinator.

Note:
1. Submission after due date will not be accepted
2. No group can change their project title and group once they submit this report.

Project Coordinator: Prof. (Dr.) Pankaj Kumar Sharma

Ms. Anshul Khanna (HOD- CS)
Ms. Mansi Mahendru

Department of CS, ABES Engineering College, Ghaziabad, Page:1

 Third Year Project Finalization

(Group No: 7 )
1. Project identification
1.1 Project title
FRAUD/ABUSE DETECTION FOR INSURANCE COMPANIES

1.2 Group members

SN Student Name Roll No
1. Rashi Sharma 2100320120137
2. Sarthak Dixit 2100320120151
3. Yuvraj Singh 2100320120199

1.3 Technical advisor(s)

1.3.1 Internal advisor:

Name: Praveen Kumar Rai
Designation: Assistant Professor
Department: Computer Science
1.3.2 Co-internal advisor (if any):
Name:

Designation:

Department:

1.3.3 External advisors (if any):

Name:

Designation:

Company/Institute:

1.4 CERTIFICATE

“This is to certify that the third year project title as stated in Sec. 1.1
proposed by student’s group mentioned in Sec. 1.2, has been found
Industry/ Software / Application/ IT / innovative (for patent) (please tick
appropriate) based and every section of this report is reflecting the same.”
(Signature of internal
advisor & date)

Department of CS, ABES Engineering College, Ghaziabad, Page:2

 Third Year Project Finalization

2. Project insights

2.1 Thematic
area(s) Research S/W Development Industry Automation
Institute Automation Social Benefits Defense Sector
Other

2.2 Keywords
Machine Learning , Data Mining , Data Analysis , Predictive Modeling ,
Convolutional Neural Networks (CNN) , Decision Tree , Neural
Network

2.3 Utilization
scope Innovative Commercial Industrial Research
Patentable Other

2.4 Execution
Plan/ Timeline Phase 1: Planning and Data Preparation
 Define project objectives, scope, and deliverables.
 Conduct a comprehensive literature review on fraud detection in
insurance using machine learning techniques.
 Identify relevant datasets and data sources for the project.
 Gather selected datasets and perform initial data exploration.
 Clean and preprocess the data to handle missing values, outliers,
and inconsistencies.
 Explore data distribution, correlations, and feature engineering
possibilities.

Phase 2: Model Development and Evaluation

 Evaluate various machine learning algorithms suitable for fraud

detection (e.g., logistic regression, decision trees, random forests).
 Develop prototype models using a subset of the data and evaluate
their performance using appropriate metrics.
 Train the selected models on the complete dataset.
 Fine-tune hyperparameters using techniques like grid search or
random search.
 Evaluate the trained models using validation datasets and cross-
validation techniques.
 Analyze model performance and iterate on feature selection and
engineering if necessary.

Department of CS, ABES Engineering College, Ghaziabad, Page:3

 Third Year Project Finalization

Phase 3: Deployment and Finalization

 Prepare the selected model(s) for deployment, including

serialization and packaging.
 Develop APIs or web services for model inference and integration
with insurance systems.
 Implement necessary security measures to protect sensitive data and
model artifacts.
 Deploy the trained model(s) to a production environment (e.g.,
cloud server, on-premise server).

2.5 S/W tools,

H/W Used Software tools:

 Programming Language: Python.

 Machine Learning Frameworks: TensorFlow, PyTorch, or scikit-
learn for model development.
 Data Processing: Pandas for data manipulation and analysis.
 Visualization: Matplotlib for data visualization.
 Version Control: Git for collaborative development and tracking
changes.
 IDE: Jupyter Notebook for interactive development and debugging.

Hardware Requirements:

 CPU: Dual-core processor (Intel Core i3 or higher) for basic model

training and inference.
 RAM: Minimum 8GB RAM for handling moderate-sized datasets
and basic model operations.
 Storage: HDD with at least 250GB of storage space for storing
datasets, models, and software.
 Networking: Basic internet connectivity for downloading datasets
and occasional updates.

Department of CS, ABES Engineering College, Ghaziabad, Page:4

 Third Year Project Finalization

3. Objective/ Scope
1. This project aims to detect fraudulent activities in insurance claims, particularly in the
healthcare sector, by utilizing data mining tools and techniques. The methodology
involves steps such as data collection, dataset preparation, developing a Gap cut
algorithm for fraud detection, training machine learning algorithms on datasets, and
deploying the model in real- life scenarios.

2. The primary goal is to enhance the efficiency of fraud detection mechanisms in

insurance management systems by leveraging advanced technologies like spectral
analysis and anomaly detection.

3. The project focuses on improving accuracy, minimizing loss, and optimizing the
process of identifying potentially fraudulent providers based on claims data.

Department of CS, ABES Engineering College, Ghaziabad, Page:5

 Third Year Project Finalization

4. Literature study
[1] Kalra, Singh, and Kumar (2022) addressed the critical issue of fraud claims detection in
insurance through the lens of machine learning. Their study likely offers insights into the
development and implementation of advanced algorithms tailored for identifying fraudulent
activities within insurance claims data, thereby contributing to the enhancement of fraud
detection mechanisms within the insurance sector.

[2] Adedotun et al. (2023) delved into the realm of predictive analytics for anticipating
automobile insurance fraud claims using machine learning methodologies. Their research
likely presents a systematic investigation into the utilization of machine learning algorithms
to forecast instances of fraudulent claims within the automotive insurance domain, thereby
aiding insurance companies in proactively mitigating potential losses.

[3] Burri et al. (2019) contributed to the field of insurance analytics by conducting a
comprehensive analysis of insurance claims using machine learning algorithms. Their study
likely offers valuable insights into the application of various machine learning techniques
for uncovering patterns, trends, and anomalies within insurance claims data, facilitating
informed decision-making processes within insurance management systems.

[4] Malhotra and Malhotra (2020) provided a detailed case study on the detection of
insurance fraud using machine learning, drawing upon real claim data for their analysis.
Their research likely offers practical implications for insurance companies by showcasing
the effectiveness of machine learning models in detecting and preventing fraudulent
activities, thereby emphasizing the importance of leveraging advanced technologies in
combating insurance fraud.

[5] Karpatne et al. (2018) explored the domain of fraud detection in insurance claims through
a data analytics approach, potentially shedding light on innovative methodologies for
analyzing and interpreting large volumes of insurance claims data to identify fraudulent
patterns. Their study likely contributes to the advancement of fraud detection techniques
within the insurance industry by leveraging data analytics methodologies to enhance the
accuracy and efficiency of fraud detection mechanisms.

[6] Sun et al. (2019) conducted a survey on the application of machine learning techniques
for detecting insurance fraud, likely providing a comprehensive overview of existing
methodologies, algorithms, and technologies employed in the detection and prevention of
fraudulent activities within the insurance sector. Their study likely serves as a valuable
resource for insurance professionals seeking to stay abreast of the latest advancements in
fraud detection techniques.

[7] Talari and Ahmad (2018) presented an innovative adaptive multi-feature approach for
insurance fraud detection, potentially introducing novel methodologies for incorporating
diverse features and variables into fraud detection algorithms. Their research likely offers
valuable insights into the development of adaptive and robust fraud detection mechanisms
capable of effectively identifying and mitigating fraudulent activities within insurance
claims data.

[8] Wang, Wang, and Ye (2019) proposed a social network analysis approach for detecting
insurance fraud, potentially introducing a novel methodology for analyzing the social

Department of CS, ABES Engineering College, Ghaziabad, Page:6

 Third Year Project Finalization

connections and relationships between different entities within insurance claims data. Their
study likely offers a unique perspective on fraud detection by leveraging social
network analysis techniques to uncover fraudulent activities within insurance networks.

[9] Shah et al. (2021) contributed to the domain of insurance fraud detection by developing
and evaluating machine learning models specifically tailored for detecting fraudulent
activities within insurance claims data. Their research likely offers practical implications for
insurance companies by showcasing the effectiveness of machine learning techniques in
identifying and mitigating insurance fraud, thereby emphasizing the importance of
leveraging advanced technologies in combating fraudulent activities within the insurance
sector.

[10] Lei et al. (2020) introduced an XGBoost-based system for financial fraud detection,
potentially offering a robust and efficient solution for detecting fraudulent activities within
financial transactions. Their research likely demonstrates the effectiveness of XGBoost, a
powerful gradient boosting algorithm, in identifying and mitigating financial fraud, thereby
contributing to the development of advanced fraud detection mechanisms within the
financial industry.

[11] Patel et al. (2019) conducted a survey paper on fraud detection and frequent pattern
matching in insurance claims using data mining techniques, potentially offering a
comprehensive overview of various data mining methodologies employed in the detection
and prevention of insurance fraud. Their study likely serves as a valuable resource for
insurance professionals seeking to understand the diverse range of data mining techniques
available for analyzing and interpreting insurance claims data to identify fraudulent patterns.

[12] Waghade (2018) conducted a comprehensive study of healthcare fraud detection based
on machine learning, potentially offering valuable insights into the application of machine
learning techniques for identifying and mitigating fraudulent activities within healthcare
insurance claims data. Their research likely contributes to the advancement of fraud
detection mechanisms within the healthcare sector by leveraging machine learning
methodologies to enhance the accuracy and efficiency of fraud detection mechanisms,
thereby safeguarding healthcare insurance systems against fraudulent activities.

Department of CS, ABES Engineering College, Ghaziabad, Page:7

 Third Year Project Finalization

References

 [1] Kalra, H., Singh, R., & Kumar, T. S. (2022). Fraud Claims Detection in
Insurance Using Machine Learning. DOI: 10.47750/pnr.2022.13.S03.053

 [2]Adedotun, F. A., Odusanya, O. A., Adesina, S. O., Adeyiga, J. A., Okagbue, H. I.,
& Oyewole, O. (2023). Prediction of automobile insurance fraud claims using
machine learning. DOI: 10.58414/SCIENTIFICTEMPER.2023.14.3.29

 [3] Burri, R. D., et al. (2019). Insurance Claim Analysis Using Machine Learning
Algorithms. International Journal of Innovative Technology and Exploring
Engineering (IJITEE), 8(8), 3986-3991.

 [4]Malhotra, R., & Malhotra, D. (2020). Detection of insurance fraud using machine
learning: A case study with real claim data. Expert Systems with Applications.

 [5]Karpatne, A., et al. (2018). Fraud detection in insurance claims: A data analytics
approach. IEEE Transactions on Big Data.

 [6] Sun, J., et al. (2019). Detecting Insurance Fraud Using Machine Learning
Techniques: A Survey. Paper presented at the 2019 IEEE International Conference
on Big Data (Big Data).

 [7] Talari, S., & Ahmad, S. (2018). An Adaptive Multi-Feature Approach for
Insurance Fraud Detection. Paper presented at the 2018 IEEE International
Conference on Big Data (Big Data).

 [8] Wang, L., Wang, Y., & Ye, L. (2019). A Social Network Analysis Approach to
Detect Insurance Fraud. IEEE Access, 7, 94758-94768.

 [9]Shah, S., et al. (2021). Insurance Fraud Detection using Machine Learning.
International Research Journal of Engineering and Technology (IRJET), 8(6), 2171-
2175.

 [10]Lei, S., et al. (2020). An XGBoost Based System for Financial Fraud Detection.
E3S Web of Conferences.

 [11]Patel, P., et al. (2019). A Survey Paper on Fraud Detection and Frequent Pattern
Matching in Insurance claims using Data Mining Techniques. International Research
Journal of Engineering and Technology (IRJET), 6(9), 3481-3485.

 [12]Waghade, S. S. (2018). A Comprehensive Study of Healthcare Fraud Detection

based on Machine Learning. International Journal of Applied Engineering Research,
13(11), 8773-8780.

Department of CS, ABES Engineering College, Ghaziabad, Page:8

 Third Year Project Finalization

Department of CS-ABES Engineering College, Ghaziabad, Page:9