Personalized Intelligent Recommendation of English Vocabulary Learning Based on Weight Kernel Learning-Support Vector Regression

The essential of vocabulary knowledge as one of the most difficult aspects of learning a foreign language is widely recognized. However, there has been a lack to analysis on English vocabulary that concentrated on authority of content and quality of Wikipedia data, such as content authority, which affect classification accuracy. This paper, proposed Weight Multiple Kernel Learning -Support Vector Regression (WMK-SVR) technique is performed efficiently in classification for enhance the accuracy. The WMK-SVR are used to transform input data into a higher dimensional space where linear regression and capture different aspects of the data, enhancing the model learning capability. wiki dataset and pre-processing involved Square Root Cubature Kalman Filter (SRCKF) is enable to cleaning the data from the document words in the English text. Feature extraction using Local Maximum Synchro Squeezing Transform (LMST) is extract the feature and offers better frequency resolution compared to conventional wavelet transforms. The proposed WMK-SVR method is evaluated using the Wikipedia data achieving a higher accuracy of 91.25% respectively. The existing techniques such as Neural Machine Translation (NMT)and Natural Language Processing (NLP) are evaluated the proposed method.

I. Introduction

English Vocabulary Learning is worldwide great significance to many non-native English speakers is foundation of language learning and it is improving the ability to speak, read and write [1]. There are two type of people English and non-English speaker which leads to performance gap in non-English languages and negatively affect user experiences. This includes longer response times, shorter context lengths and difficulties for higher-level professional. Expanding the tokenizer vocabulary by introduces frequently used long words as additional tokens help mitigate these issues [2–3]. In Machine Learning is attempted by developing structure of English vocabulary learning. However, developing such as structures was labour intensive and does not achieve acceptable performance or required domain restriction [4–5]. To make vocabulary selection and teaching more systematic, the English lexicon has been categorized. It is argued that high coverage and essential words should be the focus, particularly for beginner English language learners, who concentrate on English word families [6]. The ML based method tackles these issues by dynamically assigning weights to various modalities depend on their task relevance selecting the most beneficial and allow to capture the gap of text in efficiently achieve better outcome in English vocabulary [7].