Semantic Analysis of Public Health Medical Issues Based on Convolution Neural Networks
Algorithm 1
Data preprocessing in convolutional neural networks.
Input: training set in the form of a question
Output: numerical vector form training set
Begin
(1)
Read the dataset
(2)
Get the question part of the dataset and clear the non-Chinese character content in the question
(3)
Count all the different words in the question section and establish a Chinese character collection dictionary dicKey according to the order in which they appear
(4)
For all questions, get the maximum dimension N of the question.
(5)
Defines a full zero matrix X; the number of matrix rows is the number of problem set rows and the number of matrix columns is the maximum dimensionality N
(6)
For all questions, according to the text and the created dictionary, the question text is converted into vectors and stored in the matrix X
(7)
The training set of the numerical vector matrix is obtained
