UK Data Archive Keyword Indexing with a SKOS Version of HASSET Thesaurus

We show the evaluation results, tools and techniques used to automatically index data collections. We examine the efficiency and the accuracy of keyword automation. We tested the capacity and quality of automatic indexing using a controlled vocabulary called HASSET. We began by applying SKOS to HASSET. The automatic indexing, using the SKOS version of HASSET, provided a ranked list of candidate keywords to the human expert for final decision-making. The accuracy or effectiveness of the automatic indexing was measured by the degree of overlap between the automated indexing decisions and those originally made by the human indexer. We investigated text mining techniques to automatically index the data collection. These included applying the tf.idf model and Keyphrase Extraction Algorithm (KEA) in a Java development environment. We used Machine Learning and Natural Language Processing tools. The tools were used to build a classifier model using training documents with known keywords and then used the model to find keywords in new documents. Extensive manual and automatic evaluation was performed to calculate recall and precision scores. This poster explains how and why we applied the chosen technical solutions, and how we intend to take forward any lessons learned from this work in the future