Tid: 18 mars 2014 kl 15.15-16.00.Seminarierummet 3733, Institutionen för matematik, KTH, Lindstedtsvägen 25, plan 7. Karta!
Föredragshållare: Jesper Loso
Titel: Forecasting of Self-Rated Health Using Hidden Markov Algorithm (Master Thesis)
Abstract In this thesis a model for predicting a personís monthly average of self-rated health the following month was developed. It was based on statistics from a form constructed by HealthWatch. The model used is a Hidden Markov Algorithm based on Hidden Markov Models where the hidden part is the future value of self-rated health. The emissions were based on five of the eleven questions that make the HealthWatch form. The questions are answered on a scale from zero to one hundred. The model predicts in which of three intervals of SRH the responder most likely will answer on average during the following month. The final model has an accuracy of 80 %.
|Sidansvarig: Filip Lindskog