Skip to main content

Table 2 Linear regression of the subgroups of each class

From: Natural language processing (NLP) to facilitate abstract review in medical research: the application of BioBERT to exploring the 20-year use of NLP in medical research

 

Average annual growth rate

Proportion

Coefficient for the year of publication (95% CI)

p-value

Adjusted R2

Adjusted R2 (model with year squared)

All

20.99%

100%

72.01 (56.8–78.3)

 < 0.001

0.83

0.93

Primary research fields

 Certain infectious or parasitic diseases

42.84%

9.84%

3.46 (1.92–4.90)

 < 0.001

0.514

0.697

 Mental, Behavioral, or Neurodevelopmental disorders

39.28%

16.27%

6.23 (4.28–8.19)

 < 0.001

0.686

0.932

 Neoplasm

42.41%

27.66%

9.8377 (7.498–12.18)

 < 0.001

0.793

0.955

 Diseases of the Circulatory System

49.04%

11.27%

3.78 (2.918–4.648)

 < 0.001

0.805

0.912

Text sources

 Electronic Medical/Health and Similar databases

25.41%

57.12%

30.87 (23.33–38.41)

 < 0.001

0.784

0.931

 Published Medical Evidence

31.52%

33.84%

13.07 (11.247–14.9)

 < 0.001

0.918

0.935

 Social media + Website

51.03%

4.13%

3.133 (1.86–4.41)

 < 0.001

0.56

0.874

 Omics databases

65.08%

4.91%

1.66 (1.07–2.25)

 < 0.001

0.627

0.637

Context of use

 Bioinformatics

69.65%

15.66%

5.26 (4.23–6.29)

 < 0.001

0.849

0.853

 Clinical Decision Support and Similar fields

32.12%

25.45%

19.25 (13.42–25.09)

 < 0.001

0.7

0.94

 NLP Method Advancement

31.76%

10.47%

7.09 (4.83–9.35)

 < 0.001

0.678

0.896

 Other Medical Fields

19.44%

48.42%

21.72 (17.486–25.96)

 < 0.001

0.851

0.888

  1. NLP; natural language processing