Table 1 Definitions of commonly used terms

Accuracy: the proportion of correctly classified records:

\( \frac{\left(\mathrm{TP}+\mathrm{TN}\right)}{\left(\mathrm{TP}+\mathrm{FP}+\mathrm{TN}+\mathrm{FN}\right)} \)

False negatives (FNs): the number of records incorrectly classified as excludes. Also referred to as “missed studies.”

False positives (FPs): the number of records incorrectly classified as includes.

Prediction: a forecast of whether a record is relevant (include) or irrelevant (exclude) for a given systematic review.

Semi-automated screening tool: any web-based application that employs a combination of text mining and text classification to assist systematic reviewers during the title and abstract screening process.

Sensitivity: the ability of a screening tool to correctly classify relevant records as includes: \( \frac{\mathrm{TP}}{\left(\mathrm{TP}+\mathrm{FN}\right)} \)

Specificity: the ability of a screening tool to correctly classify irrelevant records as excludes: \( \frac{\mathrm{TN}}{\left(\mathrm{TN}+\mathrm{FP}\right)} \)

Text classification: a standard machine-learning process in which the aim is to categorize texts into groups of interest [18].

Text mining: the process of discovering knowledge and structure from unstructured data.

True negatives (TNs): the number of records correctly identified as excludes.

True positives (TPs): the number of records correctly identified as includes.