A Graph Based Framework for Removing Outliers in Time Series Boolean Signals: A Case Study of an Automatic Door

A dataset is a collection of data items gathered from various but related sources. A dataset's instance is known as a data record and can be characterised by several data attributes known as features. Today, data has become a fuel in everyday life as well as for artificial intelligence-based models for decision-making, learning, and automation. Many devices, such as sensors, cameras, and Internet of Things (IoT) devices, can be used to collect data. If there are any inconsistencies in the collected data, the outputs of traditional algorithms or models based on artificial intelligence could be negatively affected. In this paper, we have proposed a novel transition-based graph model for handling outliers in time series boolean signals. Our proposed model detects and removes the outliers from the time series boolean signals. We have simulated time series boolean signals for an automatic door and verified that our proposed model can successfully remove the outliers. Our proposed model can be used to remove outliers in situations where predictive maintenance will be carried out based on time series boolean signals,