Association Rules and Semantic Relatedness (ARSR) - Evaluation Data

<p><em><strong>ARSR</strong></em> stands for &quot;<strong><em>Association Rules and Semantic Relatedness</em></strong>&quot;,<br> a recommender system that combines association rule mining and semantic relatedness to create recommendations for form fields.</p> <p>ARSR was evaluated with focus on recommending values for fields of metadata forms.<br> The evaluation was performed with two sets of association rules (R1 and R2).<br> <strong>R1</strong> was primarily used to assess the recommendation performance.<br> <strong>R2</strong> served as an alternative and led to similar results.</p> <p>This dataset contains the raw data (&quot;<strong>collected</strong>&quot;, JSON format), generated during the evaluation.<br> It includes i.a. input (populated fields and target field), expected output, and the top 40 generated recommendations for each test combination.</p> <p>In addition, the processed data (&quot;<strong>analysed</strong>&quot;, CSV format) is provided.<br> It is based on the raw data and is used to calculate metrics and plot the results.</p> <p>The <strong>source code</strong> for ARSR and the evaluation is available on <a href="https://gitlab.com/dlr-dw/arsr">GitLab (gitlab.com)</a>.</p> <p>&nbsp;</p> <p><em>Note:</em><br> To perform the analysis on the raw data (collected) yourself, make sure to follow the setup instruction in the evaluation repository first.<br> More specifically, install dependencies and unzip &quot;data/cedar/test-instances.zip&quot; so that URI mappings can be accessed.<br> Then follow the instructions provided by the README file in the raw data archives (e.g. collected-R1.zip).</p>