Skip to main content

Table 1 Data description during preprocessing steps

From: How to automatically turn patient experience free-text responses into actionable insights: a natural language programming (NLP) approach

Hospital

Questiona

Total no of questions answered

Average no of words per answer

Original corpus size

Corpus size after pre-processing

Optimal no of topics for topic model

No of n-grams

1

Q1: remarkably well

20,982

9.13

195,579

1158

64

165

1

Q2: not as well

17,682

17.85

311,345

1814

63

117

2

Q1: remarkably well

2608

8.33

21,727

216

59

116

2

Q2: not as well

2537

24.93

63,262

628

50

119

  1. a Q1: What went remarkably well during your stay? Q2: What did not go as well during your stay?