[Retracted] Extraction and Analysis of Influencing Factors of Scientific and Technological Ability Improvement of University Teachers Based on Deep Learning Model

Algorithm 2

Steps for establishing the vector model of teacher’s secondary label.

Name: Teacher secondary label vector model establishment steps

Input: a collection of teacher’s student comments

Output: Keyword: Set of keywords

Methods and steps:

Step1: Text feature extraction

  Word2Vec eigenvalue extraction algorithm is used to extract text features from the text set of students’ comments on teaching, and the text feature vector space is constructed, among which, represents the number of input text sets.

Step2: Generate text label vector

  The text feature vector obtained in the first step is input into the algorithm of LDA model to calculate the classification probability of teachers on 30 second-level teaching evaluation labels, and the probability vector on the classification of second-level labels is obtained. The classification probability calculation formula of the document is shown in Formula (1).

(1)

(2)

  In Formula (1), represents the second-level teaching evaluation label, and the probability of in the text feature vector space of the teacher, that is, the proportion of times of in the teacher’s teaching evaluation texts of all students. In Formula (2), represents the frequency of in a particular second-level evaluation label , and represents the number of in all evaluation texts contained in the th second-level evaluation label. So in case you cannot compute if the denominator is 0, you have to add 1 to both the numerator and the denominator.

Step3: The calculation of label vector of secondary evaluation

  The final second-level evaluation label vector can be obtained by simply adding all the word vectors of the 30 second-level evaluation labels obtained in the second step.