A SHORT VIDEO POPULARITY PREDICTION USING IOT AND DEEP LEARNING

Abstract

The rapid growth of social media platforms and short-video applications has significantly increased the demand for intelligent systems that can predict video popularity in advance. Understanding the factors that influence video engagement, such as views, likes, shares, and comments, is essential for content creators, marketers, and platform providers. This project, “Short Video Popularity Prediction Using IoT and Deep Learning,” proposes an advanced framework that leverages Internet of Things (IoT) data and Deep Learning techniques to accurately predict the popularity of short videos before or shortly after publication. The proposed system integrates data collected from IoT-enabled devices such as smartphones, wearables, and user interaction sensors, which capture contextual information including user behavior, location, device usage patterns, and real-time engagement metrics. This data is combined with video-specific features such as duration, content type, visual quality, hashtags, and posting time. The system utilizes deep learning models such as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) to extract spatial and temporal features from video content and user interaction sequences. These models are capable of learning complex nonlinear relationships between input features and popularity outcomes. Additionally, attention mechanisms may be incorporated to improve prediction accuracy by focusing on the most relevant features. The system is evaluated using performance metrics such as accuracy, mean squared error (MSE), and F1-score. The results demonstrate that integrating IoT data with deep learning significantly improves prediction accuracy compared to traditional approaches. The proposed framework enables content creators to optimize their videos for maximum reach and engagement, while platforms can enhance recommendation systems. This research contributes to the fields of social media analytics, IoTbased data intelligence, and deep learning, providing a scalable and efficient solution for predicting short video popularity.