Live Stock Price Detection using ARMA-ARIMA Model with RNN Classifier

Stock market prices show characteristics similar to those of fractals, according to the Fractal Market Hypothesis (FMH), and these characteristics are attributed to varying time horizons and investor information. This occurs because every trader is relying on the same information. Main challenge focused here is Non-stationarity Data. A common requirement of used models is that data should be stationary, but in real-world applications changing statistical properties are often present. Seasonality and trends: Business environment that contains time series data always represents seasonal variations or long-term trends. Autoregressive Moving Averages (ARMAs) and Autoregressive Integrated Moving Averages (ARIMAs) have been developed for stock markets where the market becomes unstable and investors exchange information. Since the outcome of the preceding movement is taken as input conviction to the relevant step, the proposed method uses the Recurrent Neural Network (RNN) for classifying a particular time sequence. This study combines ARMA with ARIMA to model daily stock indexes on the NASDAQ stock exchange for 2022 and 2023. In the proposed work, the examine can be used to quote the upcoming months of the year and show profit and loss for the stock market over the year. NASDAQ stock exchange is a non-fractal market and an efficient market according to the proposed method to predict it.