We see numerous attempts by the research community in the past few years to make the computer capable of understanding the Natural language of Humans. The advent of Machine learning & Deep learning gave a boost to this process and resulted in successful researches. However, these Deep learning models have their own downsides and still have some negative aspects. Recent developments in Subject - Verb Agreement research using LSTM models, RNN grammar etc. have shown good results but they lag as the complexity of the sentence increases. Another approach is given by Grammarly where they use a simple and efficient GEC sequence tagger using a Transformer encoder. Their system is pre-trained on synthetic data and then fine-tuned in two stages: first on errorful corpora, and second on a combination of errorful and error-free parallel corpora. In contrary to these approaches in this paper we have tried to focus on solving this problem using Rule based Methodology with the inspiration that deterministic methodologies are the best to solve deterministic problems. Using Rule based methodology to an extent where we have carefully defined rules in English Grammar helps us to generate good results with good accuracy of 81.5% F- Scores as follow F0.5:0.8311, F1:0.855, F2 : 0.88 and with lesser resources unlike used in learning models. However, these algorithms of rule-based approach can also be equipped with language models to solve problems which have ambiguity and really have a requirement of language modelling and learning models to increase the accuracy further and give improved results.

Subject-Verb Agreement, Syntactical Error detection, Rule-based Approach