Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{187181,
        author = {Jayamurugan.R and Sindhuja.M and Ramya.M and Naveen Kumar.K and Snega.M and Sharveash Risihan.N and Alle Venkata Sai Reddy},
        title = {SMART IOT-ENABLED MILK QUALITY MONITORING SYSTEM FOR NUTRIENT AND ADULTERATION DETECTION},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {6},
        pages = {4837-4850},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=187181},
        abstract = {Milk adulteration [1][2] and contamination pose significant health risks and economic challenges within the dairy industry, necessitating rapid, accurate, and scalable quality assessment methods. Conventional laboratory-based techniques—such as lactometer tests, Gerber fat analysis, and spectroscopic assays—are precise but inherently time-consuming, costly, and reliant on specialized equipment and personnel. To address these limitations, this study proposes an integrated Internet of Things (IoT) and Machine learning [4][5] (ML) framework for real-time milk quality monitoring and Adulteration detection [1][6]. 
Our system combines an ESP8266 [2][3][5] (NodeMCU [2][3][5]) microcontroller and Arduino Uno R3 [2][3] with a suite of Sensors [1][2][3]—including pH, temperature, color, and turbidity—to capture critical physicochemical parameters. Sensor readings are preprocessed (handling missing values, outliers, and deriving milk Density [1][3][5] via an empirical equation) and transmitted via Wi-Fi to Firebase cloud [6][8], enabling continuous remote data storage and retrieval. We assembled a Dataset [17] of 1,000 samples from Kaggle, partitioned into 80% for training and 20% for testing. Five supervised ML classifiers—LightGBM [5][8], Random Forest [5][8], Decision Tree, Support Vector Machine (SVM), and Logistic Regression—were trained on standardized feature vectors. Model performance was evaluated using Accuracy [5][8], precision, recall, F1-score, and confusion matrices. 
Experimental results demonstrate that ensemble tree-based models excel in this application. LightGBM [5][8] and Random Forest [5][8] both achieved 98% Accuracy [5][8], with precision and recall exceeding 0.97 and F1-scores of 0.98. The Decision Tree yielded 97% Accuracy [5][8], while SVM and Logistic Regression attained 77% and 61% Accuracy [5][8], respectively, due to their limited handling of non-linear feature interactions. Feature importance analysis identified pH, temperature, and the derived Density [1][3][5] metric as the most influential predictors, underscoring the value of targeted Feature engineering [5][6]. 
By enabling automated, low-latency classification on low-power embedded hardware and providing an intuitive OLED [6] display and web-based dashboard for real-time feedback, our system significantly advances the state of dairy quality control. Future work will explore Deep learning [7] architectures for temporal spoilage prediction and Blockchain [10][16] integration for enhanced traceability and data integrity.},
        keywords = {Milk Quality Monitoring; IoT; Machine learning [4][5]; LightGBM [5][8]; Firebase; ESP8266 [2][3][5]; Real-Time Adulteration detection [1][6].},
        month = {November},
        }