This email is in regards to seeking mentorship for my Research Methods project and introducing my subject of interest to you all.
my Research Methods project and introducing my subject of interest to you all.
I am currently pursuing Masters in Analytics at AUT, and
I am currently pursuing Masters in Analytics at AUT, and have more than 20 years of professional experience in the information and communication technologies domain. My expertise lies in architecting solutions in telecom domain (OSS/BSS), that help customers automate and optimise their business processes, thereby minimise risks and reduce overhead costs.
As part of my research paper, I would like to explore application of Machine Learning algorithms to determine fraudulence of an insurance claim . I would be using data related to insurance claims in Automobile sector for this exercise. The framework for prediction of fraud and pattern extraction will be useful for the insurance companies in reducing costs and minimising losses.
The objective is to create an analytical and modelling framework to predict fraud in auto insurance claims based on information like demographic, policy, claim and vehicle related features in the dataset. The paper would highlight different processes or steps involved in building an efficient framework to make prediction and compare results from different classification algorithms based on defined evaluation metrics.
The research would use the data set to highlight following aspects for building the modelling framework:
Data Importing and merging
Data Imputation for missing values
Sampling of data for training, validation and testing
Feature Analysis using different visualisations
Correlation of numerical and categorical features
Feature Selection and Generation of new features
Modelling algorithms and Evaluation metrics
Hyper-parameter tuning for different algorithms
Comparison of results from different classification algorithms
I intend to use “R” software for all data exploration, analysis, visualisations and application of machine learning algorithms. The data set being used is for a prominent general insurance company, that has experienced spur in number of fraudulent claims. It comprises of 40 features (independent variables) providing information related to:
Demographics like Country, Insurer’s age, gender, education level etc.
Policy Information like coverage, limit, period etc.
Claim information like Date and Type of Incident etc.
Details about the vehicle connected to policy
Appreciate your kind feedback and guidance.
The post my Research Methods project appeared first on My Assignment Online.
