Table 12 Classification performance evaluation table before pretreatment.

From: Analyzing the capability description of testing institution in Chinese phrase using a joint approach of semi-supervised K-Means clustering and BERT

Label-Name

Label-ID

Precision

Recall

F1-score

Textile products

Label-0

\(0.8973 \pm 0.0098\)

\(0.9054 \pm 0.0014\)

\(0.9013 \pm 0.0028\)

Information Technology

Label-1

\(0.9417 \pm 0.0094\)

\(0.9598 \pm 0.0043\)

\({\textbf {0.9507}} \pm {\textbf {0.0042}}\)

Medicine and Healthcare

Label-2

\(0.8904 \pm 0.0174\)

\(0.8142 \pm 0.0082\)

\(0.8506 \pm 0.0020\)

Household Goods

Label-3

\(0.7674 \pm 0.0045\)

\(0.7820 \pm 0.0019\)

\(0.7746 \pm 0.0032\)

Agrochemicals

Label-4

\(0.8085 \pm 0.0098\)

\(0.8796 \pm 0.0059\)

\(0.8426 \pm 0.0125\)