Automatic Scoring of Spoken Language Based on Basic Deep Learning

Table 1

Summary of features.

Feature category	Feature name	Brief description of characteristics
Phonetic class	articulationRate	Speed of speech
	numSilence	Number of voice pauses
	posteriorScore	Number of pronunciation pauses
	speakingRatio	Posterior probability score of pronunciation
Text class	eassyLength	Total number of words in text
	uniqueWords	Number of nonrepeating words in the text
	parseTreeDepth	Sum of all syntactic tree depths in text
	semanticSimilarity	Semantic similarity between text and theme
	goodGrammerRatio	Correct rate of text grammar