Refrigeration showcase is utilized in super markets and convenience stores to keep various foods and drinks cool. The system, however, is susceptible to unusual events such as leakage of refrigerant coolant and frost formation, which can lead to product spoilage inside the showcases; therefore, symptoms of the unusual conditions must be identified as quickly as possible. This paper investigates possibility to apply classification methods for “showcase fault classification, assessing automatic identification by unsupervised learning.