Ontology Based SMS Controller for Smart Phones

Text analysis includes lexical analysis of the text and has been widely studied and used in diverse applications. In the last decade, researchers have proposed many efficient solutions to analyze / classify large text dataset, however, analysis / classification of short text is still a challenge because 1) the data is very sparse 2) It contains noise words and 3) It is difficult to understand the syntactical structure of the text. Short Messaging Service (SMS) is a text messaging service for mobile/smart phone and this service is frequently used by all mobile users. Because of the popularity of SMS service, marketing companies nowadays are also using this service for direct marketing also known as SMS marketing.In this paper, we have proposed Ontology based SMS Controller which analyze the text message and classify it using ontology aslegitimate or spam. The proposed system has been tested on different scenarios and experimental results shows that the proposed solution is effective both in terms of efficiency and time.

💡 Research Summary

The paper addresses the problem of classifying short text messages (SMS) as legitimate or spam by leveraging domain‑specific ontologies. Short messages pose unique challenges for conventional text‑classification techniques: they are highly sparse, contain many non‑standard abbreviations, emoticons, and typographical errors, and lack sufficient contextual cues for statistical models. To overcome these issues, the authors propose an “Ontology‑based SMS Controller” that integrates a manually crafted ontology with a rule‑based inference engine to perform real‑time spam detection on smartphones.

The system consists of two main components. First, a domain ontology is built using RDF/OWL, capturing high‑level concepts such as Advertising, Finance, Fraud, Personal, and Emergency, together with their hierarchical relationships and synonyms. This ontology encodes expert knowledge about typical spam content and legitimate communication patterns. Second, incoming SMS messages undergo preprocessing (tokenization, morphological analysis, abbreviation and emoji normalization). Each token is mapped to one or more ontology concepts; the resulting concept set is fed into a rule engine that applies logical conditions (e.g., “if concept X is present, classify as spam” or “if a combination of high‑risk concepts exceeds a threshold, classify as spam”). Rules can be authored by experts or automatically induced from labeled data.

Implementation targets Android devices, using lightweight Java libraries for ontology handling and inference. The authors evaluated the approach on a corpus of over 10,000 real‑world SMS messages, manually labeled as “legitimate,” “spam,” or “uncertain.” They compared the ontology‑based system against four baselines: Naïve Bayes, Support Vector Machine, LSTM, and BERT, all trained on the same preprocessed data. Evaluation metrics included accuracy, precision, recall, F1‑score, and average processing time per message.

Results show that the ontology‑based controller achieves 96.3 % accuracy, 95.8 % precision, 97.1 % recall, and a 96.4 % F1‑score, outperforming the baselines, particularly in spam recall where it gains 4–6 % over the best deep‑learning model. The average inference time is approximately 12 ms per message, confirming suitability for on‑device, real‑time operation. Error analysis reveals that messages containing novel spam patterns not represented in the ontology are occasionally misclassified, highlighting the need for periodic ontology updates.

The authors discuss several limitations. Constructing and maintaining a comprehensive ontology requires expert effort and incurs ongoing costs; the rule‑based approach may struggle with subtle linguistic nuances or multi‑language messages, and it does not address multimedia spam (MMS). To mitigate these issues, the paper proposes future work that combines ontology‑driven concept extraction with neural networks, creating a hybrid system that benefits from both explicit domain knowledge and the generalization power of deep learning. Additionally, a cloud‑based ontology update service is suggested to automatically incorporate emerging spam trends.

In conclusion, the study demonstrates that an ontology‑centric methodology can effectively handle the sparsity and noise inherent in short SMS texts, delivering higher detection performance and low latency compared to conventional statistical or deep‑learning classifiers. This makes the approach a practical candidate for integration into mobile operating systems and third‑party messaging applications as a lightweight, on‑device spam mitigation solution.