Deep Continuous-Time Models in Nigerian Stock Exchange Sector

Abstract

An ensemble continuous time model for predicting Nigerian stock market
is proposed in this paper. The proposed technique is the integration of
different continuous time models such as Stochastic Differential
Equation (SDE), Geometric Brownian Motion (GBM) and Constant
Elastic Variance (CEV) with Recurrent Neural Network (RNN). To
validate the effectiveness and robustness of the method, it was tested
using data from eleven sectors of Nigerian stock exchange which
comprises of closing price stock from January 2017 to December 2019.
The past returns were also tested using the aforementioned datasets, and
the results show that past returns have predictive power in predicting
future performance using Welch and F-statistics at 95% confidence
intervals. The results indicated that Deep Continuous-time model is a
promising algorithm that have the capacity to effectively predict the
Nigerian stock market