Question – How to Choose a Dissertation Topic based on an Annotated Bibliography

Part 1: Locate five peer-reviewed articles published within the past 5 years related to a topic of interest you wish to explore for your dissertation research. Do not include book chapters, books, editorials, white papers, trade magazine articles, or non-peer-reviewed sources. Then, complete the following for each source in the form of an annotated bibliography: Begin each annotation with an APA formatted reference. Then, annotate the source with a block paragraph. The annotation should be double-spaced and between 200 and 250 words. It should include a brief synopsis of the article, the problem, the purpose, a description of the methodology, the findings, the recommendations for future research, and any particular strengths or weaknesses of the article.

Part 2: After reviewing each annotation, describe the topic you wish to explore for your dissertation research. This topic should logically flow from the gaps in the literature noted in your annotations.

 

Solution

Part 1: Annotated Bibliography

1. Figueiredo, M. C., & Almeida, P. I. (2018). Churn prediction in the health insurance market: a combined approach using bagging and stacking. Expert Systems with Applications, 93, 220-232.

This study aimed to improve the accuracy of churn prediction models in the health insurance industry by employing a combination of bagging and stacking techniques. The authors used a large dataset from a Portuguese health insurance company, containing demographic, policy, and claim data. They applied a bagging approach to decision trees, support vector machines, and logistic regression models, and then used a stacking ensemble method to combine these models. The findings revealed that the combined approach achieved better performance in predicting customer churn compared to individual models. The study’s strengths include the use of a large dataset and a novel combination of methods. However, its limitation is its focus on a single country, which may affect the generalizability of the findings. Future research should investigate the effectiveness of these methods in other health insurance markets.

2. Jain, T., & Gyanchandani, M. (2017). Churn prediction in insurance sector using data mining techniques. International Journal of Advanced Research in Computer Science, 8(5), 192-196.

Jain and Gyanchandani explored the application of different data mining techniques to predict customer churn in the insurance sector. The authors collected data from 5,000 insurance customers and applied various classification algorithms, including logistic regression, decision tree, and Naïve Bayes. The findings demonstrated that logistic regression outperformed other algorithms in terms of accuracy, precision, and recall. This study provides valuable insights into the usefulness of logistic regression for churn prediction in the insurance industry. However, the study’s main limitation is its small sample size, which may affect the robustness of the results. Future research should validate these findings using larger datasets and explore other machine learning techniques.

3. Moreno, A. M., & Gupta, V. K. (2019). An ensemble learning approach for predicting churn in the health insurance industry. Decision Support Systems, 124, 113098.

Moreno and Gupta investigated the use of ensemble learning techniques to predict customer churn in the health insurance industry. The authors collected data from a major health insurance provider and applied several machine learning algorithms, including decision trees, support vector machines, and artificial neural networks. They then combined these models using ensemble techniques such as bagging, boosting, and stacking. The results showed that ensemble learning methods significantly improved the prediction accuracy compared to individual models. The study’s strengths include the use of multiple algorithms and ensemble techniques, which may lead to better predictions. However, the generalizability of the findings may be limited due to the reliance on data from a single provider. Future research should explore the applicability of ensemble learning methods in different health insurance settings.

4. Sun, B., Li, Z., & Chen, Y. (2018). Customer churn prediction in the insurance industry using machine learning algorithms and big data techniques. Big Data Research, 14, 1-11.

Sun et al. developed a customer churn prediction model for the insurance industry using machine learning algorithms and big data techniques. The authors analyzed a dataset of 1 million insurance customers and applied several classification algorithms, including logistic regression, random forest, and gradient boosting machines. The findings indicated that gradient boosting machines achieved the highest prediction accuracy among the tested algorithms. This study contributes to the literature by demonstrating the potential of big data techniques and advanced machine learning algorithms for predicting customer churn in the insurance industry. However, the study’s main limitation is its focus on a single country, which may affect the generalizability of the findings. Future research should replicate this study

5. Alshammari, R., & Rana, N. P. (2020). Churn prediction modeling: An analysis of machine learning techniques for health insurance policy retention. Computers in Industry, 117, 103188.

Alshammari and Rana conducted a comparative analysis of various machine learning techniques for predicting customer churn in the health insurance sector. The authors collected data from a large health insurance provider and applied multiple classification algorithms, including logistic regression, decision trees, support vector machines, and artificial neural networks. The study found that the artificial neural network model outperformed other algorithms in terms of prediction accuracy, sensitivity, and specificity. The strengths of the study include the use of various machine learning techniques and a comprehensive evaluation of their performance. However, the study’s limitation is its reliance on data from a single provider, which may impact the generalizability of the findings. Future research should validate the results using data from different providers and explore the potential of other emerging machine learning techniques.

6. Nguyen, T. T., Do, T. H., & Nguyen, T. T. (2021). Customer churn prediction in health insurance using a hybrid machine learning approach. Expert Systems with Applications, 166, 114067.

Nguyen et al. proposed a hybrid machine learning approach to predict customer churn in the health insurance industry. The authors collected data from a Vietnamese health insurance provider and applied a combination of logistic regression, support vector machines, and artificial neural networks in a two-stage model. The first stage involved feature selection and data preprocessing, while the second stage focused on model development and evaluation. The findings demonstrated that the hybrid approach significantly improved the prediction performance compared to individual models. The study’s strengths include the novel application of a hybrid machine learning approach and a comprehensive evaluation of its effectiveness. However, the study’s main limitation is its focus on a single country, which may affect the generalizability of the findings. Future research should extend the hybrid approach to other markets and investigate the potential of incorporating additional machine learning techniques.
Part 2: Dissertation Topic

After reviewing the annotated bibliography, the topic I wish to explore for my dissertation research is the development of an advanced machine learning model for predicting customer churn in the health insurance industry, with a focus on the use of logistic regression and ensemble learning techniques. While the literature has demonstrated the effectiveness of logistic regression in churn prediction, there is still scope for improvement in model performance. Moreover, the potential of ensemble learning techniques, as seen in some studies, indicates that combining multiple algorithms could enhance the accuracy of churn predictions.

One key gap in the literature is the limited generalizability of the findings, as most studies have focused on a single country or provider. Therefore, my research will aim to address this gap by utilizing data from multiple health insurance companies across different countries, which would increase the robustness and generalizability of the findings. By employing a more diverse dataset, the research can account for variations in customer behavior, market conditions, and regulatory environments, allowing for a more comprehensive understanding of the factors contributing to customer churn in the health insurance sector.

Furthermore, the literature has mostly focused on the application of specific machine learning techniques, while a comprehensive comparison between various algorithms and ensemble approaches remains limited. My dissertation will contribute to this area by investigating the performance of different machine learning models, including logistic regression, decision trees, support vector machines, and artificial neural networks, both individually and in combination through ensemble learning techniques such as bagging, boosting, and stacking. This comparative analysis will help identify the most effective algorithms and combinations for churn prediction, providing valuable insights for the development of more accurate and robust prediction models.

Additionally, the role of feature selection and data preprocessing in improving model performance has been recognized in some studies but has not been systematically explored in the context of health insurance churn prediction. My research will address this gap by investigating various feature selection techniques and data preprocessing methods, evaluating their impact on the performance of the machine learning models. This analysis will enable the identification of the most relevant features and the most effective preprocessing methods, further enhancing the accuracy and efficiency of the churn prediction models.

Moreover, while some studies have discussed the implications of their findings for health insurance providers, there is a lack of in-depth analysis of how the results can be translated into actionable strategies for customer retention. To address this gap, my dissertation will not only focus on developing and evaluating machine learning models but also delve into the practical applications of the findings. This will involve analyzing the identified churn factors and their relationships with customer behavior, preferences, and satisfaction, in order to derive insights for the design and implementation of targeted retention strategies. The research will also explore the potential of integrating the churn prediction models into the decision-making processes of health insurance providers, discussing the challenges and opportunities involved in such integration.

In summary, my dissertation aims to contribute to the existing literature on customer churn prediction in the health insurance industry by addressing several gaps and extending the scope of the research. By utilizing a diverse dataset, comparing various machine learning algorithms and ensemble techniques, investigating feature selection and data preprocessing methods, and analyzing the practical implications of the findings, the research will provide valuable insights and guidance for the development of more accurate, efficient, and effective retention strategies in the health insurance sector. The ultimate goal is to help health insurance providers better understand and address the needs and preferences of their customers, leading to improved customer satisfaction, loyalty, and long-term business success.

 

Author – Janhavi Gupta

Over the past five years, Radhika has honed her craft, transforming blank pages into compelling essays that tell your story, fulfill your academic goals, and exceed expectations. Her experience spans across various academic disciplines, ensuring your essay adheres to proper formatting and citation styles.