Behaviour analysis through Machine learning techniques
Sculati, Raphael (2015)
Haaga-Helia ammattikorkeakoulu
2015
All rights reserved
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2015081414000
https://urn.fi/URN:NBN:fi:amk-2015081414000
Tiivistelmä
Behaviour analysis is the science of studying the comportment of a person to establish a specific profile about it. It has firstly been used in psychology and since a few years, it has been implemented in information technology programs to improve and suggest in different forms the content of an application for users. With the growth of artificial intelligence, it tends to become the new trend that gives the possibility for applications to be personalized and centred on the user’s needs.
Machine Learning is a subcategory of artificial intelligence and has the goal to develop solutions to implement automatic methods to make our computers capable of evolving by themselves. The activities and actions of users start to be analysed to determine rules that can be integrated to align software applications in parallel with the daily routine and comportment of a person.
This thesis is part of a healthcare mobile application project (mHealth) that has for objectives to develop a management tools for the medical personal to administer the patients in the hospital. Moreover, this application would like to use the location of a user and his habits of utilisation of the software, to quickly provide information for the nurse and therefore, reduce human-machine interaction and save precious time for better purposes.
These goals are starting to be feasible through the utilisation of correct technologies and technics. This thesis analyses the different data that can be provided and uses machine learning algorithm technics to study the behaviour of a user to predict his needs and suggest him content.
Specifically, we simulate the comportment of a nurse to subsequently be construed by our machine learning solution. Thereafter, we provide the predicted content for the user via a Web Service.
The solution that we have developed has a current accuracy of 75% and the model created with simulated data will progressively adjust itself with the real data in the healthcare environment.
Machine Learning is a subcategory of artificial intelligence and has the goal to develop solutions to implement automatic methods to make our computers capable of evolving by themselves. The activities and actions of users start to be analysed to determine rules that can be integrated to align software applications in parallel with the daily routine and comportment of a person.
This thesis is part of a healthcare mobile application project (mHealth) that has for objectives to develop a management tools for the medical personal to administer the patients in the hospital. Moreover, this application would like to use the location of a user and his habits of utilisation of the software, to quickly provide information for the nurse and therefore, reduce human-machine interaction and save precious time for better purposes.
These goals are starting to be feasible through the utilisation of correct technologies and technics. This thesis analyses the different data that can be provided and uses machine learning algorithm technics to study the behaviour of a user to predict his needs and suggest him content.
Specifically, we simulate the comportment of a nurse to subsequently be construed by our machine learning solution. Thereafter, we provide the predicted content for the user via a Web Service.
The solution that we have developed has a current accuracy of 75% and the model created with simulated data will progressively adjust itself with the real data in the healthcare environment.