Evaluation Natural Language Processing Knowledge Graphs for Semi-supervised Learning

Natural Language Processing Knowledge Graphs have gained significant popularity over the last decade in numerous sectors including pharmaceutical and financial sectors. In all sectors, knowledge-based graphs showed to be popular due to their ability in providing relevant information that considers meaning and context. Hence Knowledge based graphs have provided better contextual information than other natural language processing systems. Yet knowledge graphs do not always provide complete information about the data analysed. Therefore, this work evaluated the use of knowledge graphs for entities linking and evaluation of large unstructured datasets. Thus involved the use of both semantic and linguistic modelling that when combined with deep learning enabled understanding the dataset and context of it by linking different components of the dataset together. Yet some limitations were encountered when applying these graphs in this study and there were related to computational power. Therefore, future work involves applying this work to larger datasets and with better computational power.