A Study on CET-4 Results Based on Decision Tree Algorithm

Abstract

The College English Test Band 4 (CET-4) has now become one of the most widely taken exams by college students nationwide. There are a lot of statistical data and experimental materials to prove that CET-4 has high reliability as well as high validity. It fully meets the quality requirements of a large-scale standardized test and provides an objective evaluation of the English proficiency level of the university students. It uses modern educational statistics and the scores have been equated, so it keeps the value of scores from previous years’ exams unchanged. Since the CET-4 uses a normal score system, the scores released after each test contain a great deal of information. Nowadays, with the progress of information technology and the rise of data mining, the valuable information hidden in the large amount of score data can be mined out. This paper adopts Decision Tree algorithm to analyze the CET-4 scores of Grade 2017 students from two different colleges of a university in Sichuan and it explores the CET-4 score patterns of students from two different colleges, aiming to help college teachers understand the factors affecting the test results and adjust their teaching methods, and at the same time, to make students be more clear about the direction of their future efforts, improve their learning methods and enhance their CET-4 passing rate.