Sapta Eka Putra, Yularni Putri, Faizal Burhani Ulil Fathan

Integration of AI Models and Extreme Programming for Retail Price Prediction and Inventory Optimization

Introduction

Integration of ai models and extreme programming for retail price prediction and inventory optimization. Optimize retail operations with Smart Retail AI, integrating LSTM for price prediction & XGBoost for inventory via Extreme Programming. Boost efficiency, usability, & sustainability.

Abstract

Food prices in modern retail are highly volatile and complex inventory management is often an obstacle to maintaining operational efficiency. The research developed Smart Retail AI, an artificial intelligence-based application that integrates Long Short-Term Memory (LSTM) for price prediction and Extreme Gradient Boosting (XGBoost) for stock optimization. The software development method uses the Agile Extreme Programming (XP) approach, which emphasizes rapid iteration, user engagement, and continuous testing. The test results showed that all application features worked according to the specifications through Black Box Testing, while the usability test using the System Usability Scale (SUS) resulted in an average score of 87 (Excellent category). These findings confirm that the app has high reliability and an excellent level of ease of use. The novelty of the research lies in the direct integration of AI-based predictive models into real operational retail applications with the XP cycle, thus bridging the gap between algorithmic research and practical application. Overall, Smart Retail AI contributes to improving decision-making efficiency, operational responsiveness, and business sustainability in the modern retail ecosystem.

Review

The submitted work, "Integration of AI Models and Extreme Programming for Retail Price Prediction and Inventory Optimization," addresses a critical challenge in modern retail: managing highly volatile food prices and complex inventory to maintain operational efficiency. The authors propose an innovative solution, Smart Retail AI, an application designed to enhance decision-making and operational responsiveness. This research is highly relevant given the economic pressures and competitive landscape within the retail sector, promising to contribute significantly to business sustainability by leveraging advanced technological integration. The technical approach is robust, integrating Long Short-Term Memory (LSTM) for precise price prediction and Extreme Gradient Boosting (XGBoost) for optimized stock management within the Smart Retail AI platform. The development methodology, Agile Extreme Programming (XP), is particularly noteworthy, emphasizing rapid iteration, continuous testing, and direct user engagement – factors crucial for developing practical and user-centric retail applications. The testing results presented, including successful Black Box Testing confirming functional specifications and an "Excellent" System Usability Scale (SUS) score of 87, provide initial confidence in the application's reliability and user-friendliness. The paper effectively highlights its novelty in directly integrating AI-based predictive models into a real operational retail application developed through the XP cycle, thereby bridging a significant gap between theoretical algorithmic research and practical, deployable solutions. This direct application in a production-oriented context, rather than a purely theoretical exercise, represents a valuable contribution to the field. While the abstract strongly posits improvements in decision-making efficiency and operational responsiveness, a full paper would benefit from presenting detailed performance metrics of the AI models (e.g., prediction accuracy, optimization gains) in real-world scenarios to fully substantiate the claimed operational benefits and business sustainability impact. Nonetheless, this work presents a promising framework for intelligent retail management.

Full Text

Comments

You need to be logged in to post a comment.