Deep Learning Approaches for Real-Time IoT Data Processing and Analysis
Dankan Gowda V, Praveen Damacharla, Vinod Kumar Maddineni, Venkata Akhil Kumar Gummadi · 2024
Internet of Things (IoT) device availability and its efficiency have created a large amount of data, for which real-time processing and analysis mechanisms are required. This paper focuses on the topic of deploying deep learning techniques, namely CNNs and LSTMs, to deal with certain issues of IoT data processing. A method integrating the edge and cloud computing architecture is presented to make data analysis fast and accurate. The efficiency has been proved, as in case with the CNN model which has 95% accuracy in feature extraction and the LSTM network which scored 93% in the temporal analysis, making it very efficient. The hybrid processing model summarized low latency throughout with an average of 1. it takes 5 seconds to download one figure, proving its effectiveness in real-time use. These findings also indicate that current higher architecture neural networks could be a promising candidate to improve the IoT data analysis, so it will provide a new development scope for the responsive and efficient IoT systems in multiple areas.