A novel predictive analytics model for forecasting short-term trends in equity assets prices

This paper introduces a new predictive analytics model for forecasting stock price trends in financial assets traded on major stock exchanges worldwide and the Colombian Stock Exchange. The model is built on a probability space definition that consists of a measurable space derived from filtration. In this paper, the filtration is used to index two distinct σ -algebras: one from the probability space generated by the Autoregressive Integrated Moving Average model ( ARIMA ) applied to the price of the asset and another from a probability space created by a random walk with parameters for step size and probability terms, reflecting the asset’s historical behavior. However, in other applications, different probability distribution functions can be utilized. We propose a hypothesis about the trend and assess it using the assets mentioned above. • Propose a unique approach based on a probability space definition. • Incorporate a filtration that indexes two distinct σ -algebras. • Use an autoregressive integrated moving average (ARIMA) and a parameterized random walk. • Consider 25 assets from three of the most important stock markets in the world and the Colombian stock exchange. • Demonstrate that the model produced a decisive prediction on 24 assets and no decisive prediction on only one asset.