ASSESSING AND MITIGATING FOREIGN LANGUAGE ANXIETY AMONG ENGLISH LANGUAGE LEARNERS USING A HYBRID SURVEY-ANALYTICS APPROACH

Foreign Language Anxiety (FLA) causes
students studying English as a second language to struggle
academically and lack motivation to complete their
education. This psychological barrier manifests itself in this
way since it typically leads to avoidance behaviors, reduced
involvement in classroom activities, and worse performance
outcomes. Although language education systems are fighting
FLA, research on the topic has increased. Most of the studies
now accessible lack a robust quantitative assessment
technique and do not use predictive modeling to identify at
risk children. This research intends to provide a hybrid
approach combining a machine learning categorization
model's quantitative analysis with qualitative survey tools
such the Foreign Language Classroom Anxiety Scale
(FLCAS). Two hundred fifty college students learning
English provided a dataset that had to be pre-processed.
Characteristics like communication apprehension, test
anxiety, and fear of unfavorable evaluation were drawn out
and processed through models including Logistic Regression,
Random Forest, and Support Vector Machine (SVM) to
determine the degrees of anxiety individuals experience. In
terms of accuracy, the Random Forest classifier surpassed
both the Support Vector Machine and the Logistic
Regression with an accuracy of 88.6%.