Towards New Models for the Sustainability and Development of Small Cultural Institutes

Small Cultural Institutes (SCI) have a leading role in the preservation and dissemination of cultural heritage by focusing on local societies and regional culture. Their sustainability typically depends only on their limited profits, which causes instability in their operation, especially within the current pandemic crisis. This study concerns the economic sustainability of small cultural institutes in Greece, in terms of identification and exploitation of factors that may contribute towards new sustainability models. A new original dataset was created, including several Greek museums and their economic data representing their viability status (incomes, attendance etc.). The dataset covers the period of 1998 to 2021, unfortunately including imbalanced features, which increased the difficulty of the analysis and the creation of a representative model. The study includes a statistical analysis to highlight the difficulties the museums faced during the COVID-19 pandemic. Manual preprocessing of the dataset ensured the normalization of the data to overcome biases in the training of the machine learning models. In this direction, several machine learning models have been trained to predict SCIs' sustainability with promising preliminary results. Those results show that sustainability prediction is a difficult machine learning task due to many factors; small and imbalanced datasets, few features and samples per SCI, large feature variance among SCIs, etc., are some of the most prominent.