Dynamic portfolio insurance strategy: a robust machine learning approach

Abstract

In this paper, we propose a robust genetic programming (RGP) model for a dynamic strategy of stock portfolio insurance. With portfolio insurance strategy, we divide the money in a risky asset and a risk-free asset. Our applied strategy is based on a constant proportion portfolio insurance strategy. For determining the amount for investing in the risky asset, a critical parameter is a constant risk multiplier that is calculated in our proposed model using RGP to reflect market dynamics. Our model includes four main steps: (1) Selecting the best stocks for constructing a portfolio using a density-based clustering strategy. (2) Enhancing the robustness of our proposed model with an application of the Adaptive Neuro-Fuzzy Inference Systems (ANFIS) for forecasting the future prices of the selected stocks. The findings show that using ANFIS, instead of a regular multi-layer artificial neural network improves the prediction accuracy and our model's robustness. (3) Implementing the RGP model for calculating the risk multiplier. Risk variables are used to generate equation trees for calculating the risk multiplier. (4) Determining the optimal portfolio weights of the assets using the well-known Markowitz portfolio optimization model. Experimental results show that our proposed strategy outperforms our previous model.

Can Regulation Affect the Solvency of Insurers? New Evidence from European Insurers

“…Random forest showed the highest precision, accuracy, and sensitivity Ding et al ( 2020 ) Insurance losses prediction of U.S.-based property and casualty insurance companies Four popular ML algorithms (linear regression, random forest, gradient boosting machine, and artificial neural networks) were used. Random forest showed the best accuracy and prediction Lapses Loisel et al ( 2021 ) Lapse risk management Extreme gradient boosting and support vector machine used to predict whether a policyholder will lapse her/his policy Portfolio Insurance Dehghanpour and Esfahanipour ( 2018 ) Portfolio Insurance Strategy Adaptive neuro-fuzzy inference systems (ANFIS) for prediction combined with the Markowitz portfolio optimization model for determining optimal portfolio weights Other insurance-related analysis Rawat et al ( 2021 ) Claim Analysis Two case studies were considered: (i) the health insurance sector, from the perspective of the beneficiary; and (ii) the travel insurance sector, from the perspective of the insurer Logistic regression, random forest, decision tree, support vector machine, Gaussian naïve Bayes, Bernoulli baïve Bayes, mixed naïve Bayes, K-nearest neighbors. For both case studies random forest was the best classifier with suitable feature selection methods Kang and Song ( 2018 ) Aggregate Auto-Insurance Data Analysis Feature selection techniques to classify the dataset into homogenous risk groups Rao and Pandey ( 2013 ) Factors influencing Claims in General Insurance, India Regression analysis Guelman ( 2012 ) Insurance Lost Cost Modelling Gradient boosting compared to the linear model approach …”

Application and performance of data mining techniques in stock market: A review

“…The choice of kernel function is a critical decision for prediction efficiency. From the literature reviewed, it can be inferred that the Gaussian function Chen et al, 2006;Kim, 2003 Kim and Han (2000), Kim (2006), Leigh et al (2002), Majumder and Hussian (2007), Mizuno et al (1998), Nayak et al (2014), Pan et al (2005), Roman and Jameel (1996), Schierholt and Dagli (1996), Yao et al (1999), Yao and Poh (1995), Zhang and Wu (2009) Stock price prediction (20) Alkhatib et al (2013), Arasu et al (2014), Atsalakis and Valavanis (2009), Ayodele et al (2012), Chang and Liu (2008), Choudhry and Garg (2008), Dehghanpour and Esfahanipour (2017), Hargreaves and Hao (2013), Hammad et al (2009), Hsieh et al (2011), Kim and Han (2000), Kwon and Moon (2007), Lahmiri (2011, Ou and Wang (2009), Tjung et al (2010), Tsai and Wang (2009), Vaisla and Bhatt (2010), Versace et al (2004), Wunsch et al (1998) Predicting stock markets using external factors (7) Bollen et al (2011, Kuo et al (2001), Michael et al (2005), Mittal and Goel (2012), Shriwas and Sharma (2014), Thawornwong and...…”

“…Single technique (ANN, SVM, Association rules, decision tree etc.) Hybrid techniques Hamrita andTrifi (2011), Hou et al (2013), Dehghanpour and Esfahanipour (2017) …. Versace et al (2004), Gupta and Sharma (2014) threshold strategies are used for creating trading strategies.…”

You Learn When It Hurts: Evidence in the Mutual Fund Industry

“…Recent developments in machine learning have provided new methods that simulate human learning to gain knowledge or skills or to reorganize existing knowledge to improve performance. This phenomenon has not left the mutual fund industry unaffected (e.g., Dehghanpour and Esfahanipour 2018;Zhang and Hamori 2020;Park et al 2020;Haq et al 2021). According to Chen et al (2021), as well as the remarkable advantages of machine learning, there are also limitations, such as the lack of inclusion of the prior knowledge and experience that play an important role in the learning process.…”