Résumé:
The presence of missing data in datasets poses a major challenge in data analysis, decisionmaking processes and other activities in various fields that often require specialized methods to deal with them effectively. In this paper, we propose a novel approach to dealing with missing data using models based on machine learning and deep learning, including a hybrid model with statistical and deletion methods. The proposed hybrid model leverages the strengths of Random Forest for structured data and LSTM for time-series data, providing a comprehensive solution for diverse dataset formats with varying proportions of missing data. Experimental results demonstrate the effectiveness of our approach. The hybrid RF_LSTM model achieves observation accuracy, outperforming Random Forest and LSTM, and through this work, we contribute to solving the problem of missing data by providing an efficient hybrid model that can be largely used in real-word applications
