International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 06 | Jun 2025
p-ISSN: 2395-0072
www.irjet.net
Smart Travel Advisor: Combining AI, Collaborative Filtering, and Content-Based Techniques Savitha M Assistant Professor, Dept. of Computer Science and Engineering, VCET, Puttur, Karnataka, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Modern travel platforms often fail to adapt to
interaction data and destination features to deliver tailored suggestions. Incorporating collaborative filtering, contentbased filtering, and popularity-based techniques, the system adapts to both new and returning users. AI-generated descriptions and real-time imagery enhance the recommendations, providing travellers with comprehensive and visually engaging information. The result is a dynamic, user-centric platform that transforms the travel planning experience through intelligent automation and contextual awareness.
individual user preferences, offering generalized suggestions that do not account for personal interests or behavior. This work addresses that gap by developing an AI-based tourist advisor system that delivers personalized destination suggestions using machine learning. Data is collected from user interactions and destination attributes stored in a MySQL database. The system employs collaborative filtering with SVD, content-based filtering using TF- IDF and cosine similarity, and a popularity-based approach based on click frequency. Enrichment is achieved through integration with Gemini AI for content generation and Google Custom Search for real- time imagery. This project demonstrates a complete, adaptive travel recommendation platform that enhances planning efficiency and user engagement.
2. LITERATURE SURVEY S. Sankar et.al [1] propose an intelligent travel planning system that enhances user experience by integrating Google Street View with AI-driven attraction scoring. Machine learning algorithms evaluate local destinations and overlay insights onto real-time virtual environments, enabling users to preview locations before finalizing their itineraries. This significantly reduces last-minute changes and boosts planning confidence. The study also explores how visual previews influence decision-making, highlighting the potential of immersive virtual tools in modern tourism.
Key Words: Travel Recommendation, Machine Learning, Collaborative Filtering, TF-IDF, Gemini API
1. INTRODUCTION Travel is an integral part of modern life, offering individuals opportunities for exploration, relaxation, and cultural enrichment. However, with the increasing availability of travel options and vast amounts of information, travellers often struggle to make informed decisions tailored to their personal preferences. Traditional travel platforms typically present static suggestions based on popularity or location, without adapting to individual needs, interests, or contextual factors like weather and budget. This lack of personalization leads to generic recommendations that may not align with user expectations, resulting in inefficient planning and reduced satisfaction.
R. Semwal et.al [2] explore real-time personalization in Tourism 3.0 using AI and machine learning. Their framework analyses user interactions—such as prolonged engagement with lodging or event pages—using natural language processing and behavioural modelling. The system adapts travel suggestions dynamically based on evolving interests, leading to improved user engagement with lesser-known destinations. The authors emphasize transparency and recommend incorporating explainable AI techniques for ethical implementation.
In today’s AI-driven digital landscape, personalized recommendation systems have proven effective across domains such as e-commerce and entertainment, yet their application in travel planning remains underdeveloped. Travel data presents unique challenges, combining user behavior, location attributes, seasonal trends, and dynamic pricing all of which mus be accurately interpreted to deliver relevant suggestions. Furthermore, effective recommendations must go beyond destination names, offering enriched content and visual insights to assist in realworld decision-making.
R. R. Manthena et.al [3] introduce Route Chat Connect, a collaborative travel platform built with Python’s Streamlit and the True Way Directions API. This system allows users to plan trips interactively by merging chat functionality with route mapping, enabling real-time discussions on accommodations and activities. By consolidating communication and itinerary management, it streamlines group decision-making. However, the model lacks predictive tools for price fluctuations and does not support live congestion monitoring.
This project addresses these gaps by developing a smart travel recommendation system that leverages user
B. S. S. Miryalkar et.al [4] present Journey Craft, a chatbotbased travel planner that personalizes itineraries through
© 2025, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 665