Optimizing Mobile Typing Efficiency: A Comparative Analysis of N-Gram, LSTM, and Transformer-based Next-Word Prediction Models

Nossam Shamita Reddy, P Hemalatha, P Naga Hari Priya, Narpala Vasantha, Nayaka Madhavi and G Thippanna

International Journal of Scientific Research in Science and Technology, 2026, vol. 13, issue 3, 807-813

Abstract: The proliferation of misinformation on social media platforms poses a significant threat to public discourse and societal stability. Traditional approaches to fake news detection largely rely on natural language processing (NLP) techniques that analyze the textual content of news articles. However, these methods often fail to capture the complex propagation patterns and user interaction dynamics that characterize the spread of misinformation. This paper proposes a novel framework utilizing Graph Neural Networks (GNNs) to detect misleading news by integrating semantic content features with social graph topology. Specifically, the proposed model constructs a heterogeneous graph where nodes represent news articles and social media users, and edges represent interaction behaviors. We utilize BERT transformers for high-dimensional text embedding and employ Graph Convolutional Networks (GCN) to aggregate neighborhood information. Experimental results on the FakeNewsNet dataset demonstrate that the proposed method outperforms state-of-the-art baseline models that rely solely on text or static network features, achieving a significant improvement in detection accuracy. With the ubiquity of mobile computing, virtual on-screen keyboards have become the primary input method for digital communication. However, the lack of tactile feedback and limited screen real estate often result in slow typing speeds and high error rates. This paper proposes a robust Next-Word Prediction (NWP) system designed to enhance user typing efficiency by suggesting contextually appropriate words. We implement and evaluate three distinct language modeling architectures: N-gram statistical models, Long Short-Term Memory (LSTM) networks, and a lightweight Transformer-based model. The models are trained on a hybrid dataset comprising Wikipedia entries and SMS corpora to capture both formal and conversational contexts. Experimental results demonstrate that while Transformer models achieve the lowest perplexity and highest prediction accuracy, LSTM networks offer a competitive balance between accuracy and inference latency suitable for resource-constrained mobile environments.

Keywords: Natural Language Processing; Next-Word Prediction; Language Modeling; LSTM; Transformer; Mobile Computing (search for similar items in EconPapers)
Date: 2026
References: Add references at CitEc
Citations:

Downloads: (external link)
https://ijsrst.com/home/article/view/IJSRST26133205 Abstract page (text/html)
https://ijsrst.com/home/article/download/IJSRST26133205/IJSRST26133205 Full text (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:etm:ijsrst:v13:y2026:i3:id:1669

DOI: 10.32628/IJSRST26133205

Access Statistics for this article

More articles in International Journal of Scientific Research in Science and Technology from Technoscience Academy
Bibliographic data for series maintained by Pankaj Sharma ().