A Quantitative Analysis of Women’s Health Investments

Abstract

Investments in women’s health remain significantly underrepresented despite mounting evidence of their societal and economic value. This thesis presents a data-driven framework for evaluating and quantifying the financial performance of publicly traded companies focused on women’s healthcare. The research introduces a novel financial index quantifying women’s health constructed with various features through management infrastructure. To assess the viability and impact of investing in this sector, the index is used as the basis for simulating portfolio strategies under a long-term buy-and-hold strategy. A key methodological contribution is the application of unsupervised learning, particularly KMeans and Spectral Clustering, to identify firms with similar ESG and financial profiles. These clusters reveal latent structures overlooked by traditional industry classifications and support the design of a women’s health-focused investment portfolio. To enhance interpretability, t-stochastic neighbor embedding distributed (t-SNE) is used to visualize the high-dimensional feature space and clustering outcomes. Backtesting confirms the potential for these clustered portfolios to deliver competitive, risk-adjusted returns while aligning with gender-focused investment objectives. Overall, this research advances gender-lens investing by integrating machine learning with financial modeling, offering a scalable methodology for thematic index construction and classification in the women’s healthcare sector.