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{196457,
        author = {Devansh Bhosale and Vighnesh Chorge and Adit Biramne and Harsh Aujee and Dhanashri Kane},
        title = {Multi-Source News Synthesizer using Deep Learning},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {4643-4650},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=196457},
        abstract = {The fast expansion of digital news platforms brought a vast amount of information which correspondingly led to an explosion of information’s, misinformation and disinformation. Such misinformation, which is usually politically, financially or ideologically motivated, undermines the credibility of digital journalism and informed decision making. As mentioned by Rajesh et al. (2019) there is no method of verifying web content and hence authenticity of news comes into question to a large extent. To address this issue, scholars have proposed the automatic detection of fake news using machine-learning and natural language processing approaches. These methods follow the modular pipeline to preprocess data and satisfy feature extraction & classification. The more similar the language, the better a pre-processing like (text pre-processing operations tokenization, stop word removal and stemming) might work. Other well-known methods for text feature representation are: bag-of-words, n-grams, and TF-IDF. We experimented with few ML classifiers (Naïve Bayes, Logistic Regression, SVM (Support Vector Machines), Random Forest and Stochastic Gradient Descent) and the best F1 Score is around 72% using LR (Logistic regression) with TF-IDF and N-Gram features. Studies by Campan et al. 3 concentrates on the credibility of sources and media context, and Dey et al. propose bias detection aware model for political news. These results demonstrate that a combination of linguistic, probability and context suggestions provide substantial improvements in detection performance. 6 Conclusions We find that the surveyed solutions provide strong evidence in support for ML based pipelines to eliminate or at least highly reduce human subjectivity in news verification. They also identify direction for future research for deep learning models, semantic embedding and hybrid ensemble approaches with the massive scale. However, at the same time, we can also tell from related work that linguistic clues alone might not be sufficient to capture more subtle context dependent mechanisms of misinformation and real-time fake news detection.},
        keywords = {Machine Learning, Misinformation Analysis, Fake News Detection, Machine Learning, Natural Language Processing, Text Classification Feature Extraction TF-IDF N-Grams Source Credibility Analysis Context -Aware News Verification Digital Journalism Automated News Validation.},
        month = {April},
        }