Raising environmental awareness through deep learning approach: via NLP
Abstract – Since the environmental issue is worsening, including increasing plastic usage and rising temperature, this is the first thing we need to solve now. This study aims to provide an AI-based method for assessing text-based information in light of the growing significance of the environment. The dataset was collected from the Kaggle website with the aim of experimenting. The pre-trained BERT deep learning method was utilized for converting natural language (English) into numerical values. In the classification step, the Gaussian NB was utilized since we assumed a high likelihood of a ‘curse of dimensionality’ given the characteristics of the transformed dataset (100 rows and 48 columns). In each experiment, our recommended classifier (Gaussian NB) beat other classifiers, achieving an average of 88%. These are the principal findings of this research: at first, it can be deduced that the suggested model, which performed well in text mining, may be used for various issues in addition to environmental ones. Second, because the Gaussian NB was proficient at coping with large datasets, we demonstrated that it outperformed other machine learning algorithms in text mining. Lastly, as the proposed algorithm achieved a high accuracy with the limited datasets, it could be deduced that the model would perform better if it gathered sufficient datasets. In the end, this paper concludes that analyzing environmental issues through text mining is quite successful. This result could be useful for generating environmental reports and raising environmental awareness in the future.