Application of Machine Learning in Intelligent Medical Image Diagnosis and Construction of Intelligent Service Process

Table 2

Factors and connotations.

Factor	Connotation	Explanation
Subjective	Different domain knowledge, different cognition ability of causal relationship	Doctor: in the process of diagnosis, according to the medical knowledge he has mastered and combined with the clinical symptoms of the patient, he will comprehensively reason about the disease and explain the cause and pathology in detail
		Model designers: have deep learning model design and other related algorithm knowledge, and use deep learning directly as a black box. Based on statistics, explain the parameter tuning process and output results
Cognitive	Cognitive ability is determined by perception, memory ability, etc., which puts forward requirements for the complexity and scale of the model and the expression method of the model	Linear models can understand the importance of features from the perspective of weights, but how can it be understood by a series of linear and nonlinear combinations of neural networks
		A simple decision tree model can quickly and intuitively get interpretable results, but a complex decision tree puts forward higher requirements on human perception and memory
Objective	Accurately judging and refining the causal relationship between the input and output of a deep learning model is an objective criterion for judging model interpretability	In the medical field, the diagnosis results are based on the current physical condition of the patient, the lesion characteristics of the medical imaging response, the clinical observation data and the medical knowledge of the doctor
		In the field of deep learning, how is the relationship between input samples and output results, how to quantify, how strong the relationship is, and how to approximate the causal relationship