CRACS - Book Chapters

Permanent URI for this collection

http://localhost:4000/handle/123456789/127

Browse

Browsing CRACS - Book Chapters by Author "João Gama"
  • Item
    Predicting Ramp Events with a Stream-based HMM framework
    ( 2012) João Gama ; Audun Botterud ; Vladimiro Miranda ; Vítor Santos Costa ; Carlos Ferreira
    The motivation for this work is the study and prediction of wind ramp events occurring in a large-scale wind farm located in the US Midwest. In this paper we introduce the SHRED framework, a stream-based model that continuously learns a discrete HMM model from wind power and wind speed measurements. We use a supervised learning algorithm to learn HMM parameters from discretized data, where ramp events are HMM states and discretized wind speed data are HMM observations. The discretization of the historical data is obtained by running the SAX algorithm over the rst order variations in the original signal. SHRED updates the HMM using the most recent historical data and includes a forgetting mechanism to model natural time dependence in wind patterns. To forecast ramp events we use recent wind speed forecasts and the Viterbi algorithm, that incrementally nds the most probable ramp event to occur.
  • Item
    Predictive Sequence Miner in ILP Learning
    ( 2012) Vítor Santos Costa ; Carlos Ferreira ; João Gama
    This work presents an optimized version of XMuSer, an ILP based framework suitable to explore temporal patterns available in multi-relational databases. XMuSer's main idea consists of exploiting frequent sequence mining, an efficient method to learn temporal patterns in the form of sequences. XMuSer framework efficiency is grounded on a new coding methodology for temporal data and on the use of a predictive sequence miner. The frameworks selects and map the most interesting sequential patterns into a new table, the sequence relation. In the last step of our framework, we use an ILP algorithm to learn a classification theory on the enlarged relational database that consists of the original multi-relational database and the new sequence relation. We evaluate our framework by addressing three classification problems and map each one of three different types of sequential patterns: frequent, closed or maximal.