Towards a Semantic Similarity Recommendation Strategy for Big Data Repositories

Abstract

A measurement process should be repeatable, extensible to new requirements, and comparable measurements. This is so to enable the same measurement process to be implemented by different actors so that their results are comparable over time. Once the measurement process has been defined, the measurements need to be interpreted in order to be able to make decisions and, based on this, recommend courses of action. It may happen that depending on the entity that is under monitoring, and a particular situation, there is not always background information or prior knowledge available. This would mean that, despite having identified the situation, it would not be possible to provide recommendations or courses of action configured for the specific scenario. Additionally, the Internet of Things (in English, Internet of Things - IoT) has allowed the development of data collection strategies, supporting decision-making processes in real-time. The objective is to develop a recommended strategy based on measurements, based on semantic similarity analysis that allows providing experiences and/or recommendations by analogy with another entity.