A new methodology to support wind investment decision: a combination of natural language processing and Monte Carlo option pricing technique

This paper presents an innovative approach based on the Monte Carlo option pricing model and Natural Language Processing (NLP) to support investment decisions in wind farm projects. Thanks to recent ever deeper architectures such as Transformers, the linguistic capabilities of NLP models have become increasingly advanced to the point of helping in decision-making. The investment decision in these projects can be tough considering the uncertain economic performance depending on the stochastic nature of revenue. The idea is to price the managerial flexibility of changing investment decisions during the project lifetime depending on the wind investment's profitability by treating it as a financial option. We use the Monte Carlo option pricing technique to achieve this result by letting the project value evolve through a Geometric Mean Reverting process. This technique is combined with sentiment analysis, one of the most popular NLP tasks evaluating natural language sentences. By integrating the polarity value, we can modify the transition probabilities from one phase of the investment to another and, consequently, the profitability of the investment.