This thesis investigates the prerequisites needed for the Swedish real estate company Fabege to create useful machine learning models for classification and prediction of error reports from tenants. These error reports are regarding cold indoor climates and bad indoor air quality. By analyzing the available data, that consists of error reporting data, weather data and indoor climate data, the thesis investigates the different correlations between the sensor data and the error reports. By using an algorithm called decision jungle, two machine learning models have been trained in Microsoft Azure Machine Learning Studio. The main model, trained on error reporting data and weather data, shows the possibilities to classify data instances as a part of different error reporting classes. The model proves that it is possible to predict the emergence of future error reports of different classes with an average accuracy of 78%. The complementary model, trained on a small but more richly annotated dataset consisting of one year of indoor sensor data as well as the above-mentioned data, shows that there is a possibility to improve the main model by using indoor climate data. The thesis has shown that for Fabege to expand and improve these models, the amount of data collected from the indoor sensors needs to be largely increased. Fabege also needs to improve the quality of the error reporting data, which could be achieved by improving the error reporting form used by the tenants.

Klassificering av dagliga aktiviteter (ADL) kan användas i system som bevakar människors aktiviteter i olika syften. T.ex., i nödsituationssystem. Med machine learning och bärbara sensor som samlar in data kan ADL klassificeras med hög noggrannhet. I detta arbete, ett proof-of-concept system med tre olika machine learning algoritmer utvärderas och jämförs mellan tre olika dataset, ett som är allmänt tillgängligt på (Ugulino, et al., 2012), och två som har samlats in i rapporten med hjälp av en android enhet. Algoritmerna som har använts är: Multiclass Decision Forest, Multiclass Decision Jungle and Multiclass Neural Network. Sensorerna som har använts är en accelerometer och ett gyroskop. Resultatet visar hur ett konceptuellt system kan byggas i Azure Machine Learning Studio, och hur tre olika algoritmer presterar vid klassificering av tre olika dataset. En algoritm visar högre precision vid klassning av Ugolino’s dataset, jämfört med machine learning modellen som ursprungligen används i rapporten. / Classifying Activities of daily life (ADL) can be used in a system that monitor people’s activities for different purposes. For example, in emergency systems. Machine learning is a way to classify ADL with high accuracy, using wearable sensors as an input. In this paper, a proof-of-concept system consisting of three different machine learning algorithms is evaluated and compared between tree different datasets, one publicly available at (Ugulino, et al., 2012), and two collected in this paper using an android device’s accelerometer and gyroscope sensor. The algorithms are: Multiclass Decision Forest, Multiclass Decision Jungle and Multiclass Neural Network. The two sensors used are an accelerometer and a gyroscope. The result shows how a system can be implemented using Azure Machine Learning Studio, and how three different algorithms performs when classifying three different datasets. One algorithm achieves a higher accuracy compared to the machine learning model initially used with the Ugolino data set.