A TF-IDF Symptom-Matching Framework for Tabular Knowledge-Based Tomato Disease Decision Support
DOI:
https://doi.org/10.29207/bnnejy35Keywords:
tomato disease, decision support system, TF-IDF, information retrieval, plant pathology, knowledge baseAbstract
Early and accurate field identification of tomato diseases is critical for timely intervention, yet farmers in resource-limited settings frequently lack access to expert diagnosis. This paper proposes a lightweight, knowledge-based decision support system (DSS) for differentiating nine economically important tomato diseases and pests using a recently published tabular knowledge base. The DSS operates in two stages: first, it represents each disease's symptom description as a TF-IDF vector (unigram and bigram, lemmatized, with domain-specific stopword filtering) and retrieves the best-matching disease via cosine similarity to an observed symptom query; second, it refines the ranking using a rule-based environmental condition filter (warm-humid or cool-wet) when field context is available. We evaluate the framework on 121 individual symptom sentences extracted from the knowledge base as independent field-observation queries. The symptom-similarity stage alone achieves a sentence-level top-1 retrieval accuracy of 99.2% (120/121 sentences), with only one Mosaic Virus sentence – the non-specific phrase ‘stunted growth and reduced yields’ – initially misranked. Symptom cosine similarity analysis reveals that Target Spot is the most structurally confusable disease, sharing the highest off-diagonal similarity with Early Blight (0.144), Late Blight (0.137), and Septoria (0.131). Treatment complexity analysis further shows a clinically important distinction between fungal/pest diseases (20 or more chemical treatment options each) and viral diseases (Mosaic Virus: 2; Yellow Leaf Curl Virus: 1), where chemical interventions are ineffective and prevention-only strategies are documented.
References
[1] G. Salton and C. Buckley, “Term-weighting approaches in automatic text retrieval,” Information Processing & Management, vol. 24, no. 5, pp. 513–523, 1988.
[2] Smart Agriculture Research Group, “Smart Agriculture: An Expert System for Tomato Plant Disease Diagnosis,” in Proc. Intelligent Systems and Applications, Springer, 2024.
[3] S. S. Harakannanavar, J. M. Rudagi, V. I. Puranikmath, A. Siddiqua, and R. Pramodhini, “Plant leaf disease detection using computer vision and machine learning algorithms,” Global Transitions Proceedings, vol. 3, no. 1, pp. 305–310, 2022.
[4] S. Panno et al., “A review of the most common and economically important diseases that undermine the cultivation of tomato crop in the Mediterranean basin,” Agronomy, vol. 11, no. 11, Art. 2188, 2021.
[5] K. Shah, “Development of a web-based tomato plant disease detection and diagnosis system using transfer learning techniques,” Journal of Civil Engineering and Environmental Technology, vol. 11, 2024.
[6] M. O. Omia et al., “Deep learning for crop pest and disease detection: A step-by-step guide from data collection to model deployment,” Computers and Electronics in Agriculture, vol. 223, 2024.
[7] S. Saleem, J. Iqbal, and M. H. Manzoor, “An effective TF/IDF-based text-to-text semantic similarity measure for text classification,” in Proc. International Conference on Theory and Practice of Digital Libraries, 2014, pp. 362–373.
[8] Z. Al-Qudah and R. Al-Haidari, “A review of deep learning-based plant disease detection systems: challenges and future directions,” IEEE Access, vol. 12, pp. 35401–35420, 2024.
[9] S. Patel and A. Kumar, “TF-IDF vectorization with cosine similarity for information retrieval in biomedical text,” in Proc. International Conference on Advances in Computing and Communication, 2021.
[10] K. A. Larga and M. D. Santos, “A Comprehensive Dataset of Tomato Plant Diseases for National Predictive Analytics and Crop Health Management,” IEEE DataPort, 2026, doi: 10.21227
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Journal of Systems Engineering and Information Technology (JOSEIT)
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under Creative Commons Attribution 4.0 International License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (Refer to The Effect of Open Access).
