|
|
 |
| |
|
|
|
Leveraging Deep Learning with Alpine Skiing Optimizer in Financial Performance Forecasting and Financial Statement Analysis: A Case Study from Saudi Arabia |
|
|
|
PP: 565-577 |
|
|
doi:10.18576/jsap/150314
|
|
|
|
Author(s) |
|
|
|
Manal M. Khayyat,
Babikir Mubarak O. Elsheikh,
Ahmed Edris Abdu,
Buthayna G. E. Tamim,
Alhan H. A. Awadallah,
Abdelgalal O. I. Abaker,
Nadia Bushra Mohammed Ali,
|
|
|
|
Abstract |
|
|
| Financial predicting depends upon the usage of present and past financial data to create the best forecast of the future economic condition, evade higher risk states, and upsurge benefits. Such forecasts are main to anyone who needs to get the condition of possible finances in the future, with decision-makers and investors. However, the intricate nature of financial data makes it challenging to acquire precise forecasts. Artificial intelligence (AI), which was exposed to be appropriate for analyzing very intricate issues, can be employed for financial prediction. Currently, Machine Learning (ML) scholars have come up with numerous methods and a massive amount of studies have been published consequently. As such, a major quantity of surveys exist that cover ML for financial predicting studies. In recent times, Deep Learning (DL) systems started seeming in the field, with outcomes that significantly outperform conventional ML counterparts. This study develops a Deep Learning in Financial Performance Forecasting and Financial Statement Analysis (DLFPF-FSA) model. The DLFPF-FSA model relies on improving the prediction of financial performance using state-of-the-art optimization algorithms. To accomplish that, the data normalization stage is initially applied by z-score normalization to standardize data by scaling it to a common range. Next, the feature selection process has been executed by the bald eagle search optimization (BESO) algorithm to identify the most relevant features from input data. For the prediction of financial performance, the DLFPF-FSA system designs a hybrid of long short-term memory and gated recurrent unit (LSTM+GRU) method. Eventually, the alpine skiing optimizer (ASO) adjusts the hyperparameter values of the LSTM+GRU algorithm optimally and outcomes in greater prediction performance. The experimental evaluation of the DLFPF-FSA algorithm can be tested on a benchmark dataset. The extensive outcomes highlight the significant solution of the DLFPF-FSA approach to the financial prediction process |
|
|
|
|
|
|
|
