Japanese Dependency Analysis using Multi-Kernel Support Vector Machine based on Term Frequency and Inverse Document Frequency

Nowadays, Japanese dependency analysis examines the grammatical structure of sentences and identifies the relations between words and phrases of text. This analysis is involving with natural language processing (NLP) with in dependency analysis is particularly important due to its complex sentence structure and grammatical relations. However, dependency analysis is the complexity of handling long-distance dependencies and nested clauses, which can lead to accuracy issues and increased computational requirements. To overcome those issues this research proposes an ensemble model of Multi-Kernel Support Vector Machine (MK-SVM) for Japanese dependency analysis. Initially the data is collected from the Japanese Text Dataset (JTD). The dataset is preprocessed through normalization, stemming, and stop-words removal, followed by feature extraction using Term Frequency-Inverse Document Frequency (TF-IDF) and N-grams. The preprocessed data is then classified using MK-SVM, which combines multiple kernel functions to improve accuracy. By leveraging the strengths of different kernels, MK-SVM enhances classification performance. The proposed MK-SVM model is used to overcome drawbacks of existing system terms of metrices the values are of 97.01 % of precision, 96.01 % of recall, 98.96% of F1 score and 96.82% of accuracy, when compared to the existing method like k-nearest neighbors (KNN) respectively.