This dataset contains the results of the DZHW-funded project Mining Acknowledgement Texts in Web of Science (MinAck). Please read readme.txt to know more about the files stored in this repository. The aim of the MinAck project is to conduct an analysis of acknowledgment texts from the Web of Science (WoS) using the FLAIR NLP Framework. The FLAIR NLP Framework is used to perform the acknowledged entity recognition task. Our NER model is able to recognize 6 entity types: funding agencies (FUND), corporations (COR), universities (UNI), individuals (IND), grant numbers (GRNB) and miscellaneous (MISC). The NER model was trained with the dataset containing over 600 annotated sentences from acknowledgement texts, written in scientific articles stored in WoS. Afterwards, the model was applied to analyse a corpus of approx. 200,000 acknowledgement texts from WoS. The data chosen for the present study was restricted by year and scientific discipline. Records from four different scientific disciplines published from 2014 to 2019 were considered: two disciplines from the social sciences (sociology and economics) and oceanography and computer science for comparison. Additionally, only WoS records types “article” and “review”, published in a scientific journal in English were selected.