Personalized Advertisement Recommendation System for a Social Media Platform in Tamil

With the widespread use of social media platforms, personalized advertisements have become crucial for effective marketing. This research paper presents the development of an advertisement personalization system tailored for the Tamil language on social media platforms. The system comprises four major components: keyword extraction from social media posts, sentiment analysis for comments, next word prediction for comments and posts, and tracking and extracting Tamil words from images. To enable effective targeting of advertisements, keyword extraction from social media posts is employed. The system utilizes natural language processing techniques to extract relevant keywords that reflect the interests and preferences of Tamil-speaking users. And the component of sentiment analysis plays a vital role in understanding the sentiment expressed in comments. By utilizing machine learning algorithms and sentiment lexicons specifically designed for Tamil, the system categorizes comments as positive or negative. Furthermore, Next word prediction is incorporated to enhance the engagement and efficiency of user interactions on social media platforms. By leveraging statistical language models and n-gram, the system suggests the most probable Tamil words or phrases to users as they compose comments or posts, facilitating faster and more accurate communication. Moreover, the system integrates image tracking and Tamil word extraction techniques to process visual content. Through the utilization of concurrent neural network algorithm and optical character recognition (OCR) specifically trained for Tamil script, the system extracts relevant Tamil words from images shared on social media platforms. The proposed advertisement personalization system aims to improve the effectiveness of marketing campaigns, enhance user engagement, and deliver a personalized experience for Tamil-speaking users.