#### DMCA

## Learning Routines Over Long-Term Sensor Data Using Topic Models

### Citations

4346 | Latent Dirichlet allocation
- Blei, Ng, et al.
- 2003
(Show Context)
Citation Context ...s (VB) Variational Bayes inference consists on defining a parametric family of distributions that forms a tractable approximation to an intractable true joint distribution. Blei’s original LDA model (=-=Blei et al. 2003-=-) proposed a variational Bayes ExpectationMaximization inference algorithm in order to estimate the parameters in the training phase of the algorithm. They suggest a variational Dirichlet distribution... |

1809 | A solution to Plato’s problem: The Latent Semantic Analysis theory of the acquisition, induction, and representation of knowledge
- Landauer, Dumais
- 1997
(Show Context)
Citation Context ...pic is a probability distribution over words. One of the first statistical models employed to obtain semantic information from a word-document co-occurrence matrix was Latent Semantic Analysis (LSA) (=-=Landauer & Dumais 1997-=-). In LSA, words and documents are represented as points in Euclidean space. In contrast, in topic models the relations between words and documents are expressed in terms of probabilistic topics. Hofm... |

1223 | Probabilistic latent semantic indexing
- Hofmann
(Show Context)
Citation Context ...ce. In contrast, in topic models the relations between words and documents are expressed in terms of probabilistic topics. Hofmann introduced the probabilistic Latent Semantic Indexing method (pLSI) (=-=Hofmann 1999-=-) (Hofmann 2001), which does not make any assumption about how the mixture weights are generated, making difficult to classify new documents when the model is trained. Blei, Ng & Jordan (2003) extende... |

1098 | Finding scientific topics - Griffiths, Steyvers - 2004 |

614 | Unsupervised learning by probabilistic latent semantic analysis
- Hofmann
(Show Context)
Citation Context ..., in topic models the relations between words and documents are expressed in terms of probabilistic topics. Hofmann introduced the probabilistic Latent Semantic Indexing method (pLSI) (Hofmann 1999) (=-=Hofmann 2001-=-), which does not make any assumption about how the mixture weights are generated, making difficult to classify new documents when the model is trained. Blei, Ng & Jordan (2003) extended the pLSI mode... |

494 | Unsupervised learning of human action categories using spatial-temporal words - Niebles, Wang, et al. - 2007 |

365 | 2004) “The Author-Topic Model for Authors and Documents - Rosen-Zvi, Griffiths, et al. |

207 | Online learning for Latent Dirichlet Allocation - Hoffman, Blei, et al. |

197 | Discovering object categories in image collections - Sivic, Russell, et al. - 2005 |

116 | Unsupervised activity perception in crowded and complicated scenes using hierarchical bayesian models - Wang, Ma, et al. - 2009 |

110 | Evaluation methods for topic models - Wallach, Murray, et al. - 2009 |

105 | A survey of context modelling and reasoning techniques,” Pervasive Mob - Bettini, Brdiczka, et al. - 2010 |

95 | Discovery of activity patterns using topic models. - Huynh, Fritz, et al. - 2008 |

93 | An architecture for parallel topic models
- Smola, Narayanamurthy
- 2010
(Show Context)
Citation Context ...er of latent topics Due to the amount of time required to train the LDA model with MERL data using VB inference algorithm (setting more than 50 topics), we decided to employ Yahoo LDA implementation (=-=Smola & Narayanamurthy 2010-=-). This implemen11 Table 2: Some examples of 4 most probable words for some topics obtained with a LDA model trained with VB (T = 100) using the Innotek dataset Topic 4 Topic 10 Topic 27 Topic 62 Topi... |

92 | Discovering hidden time patterns in behavior: T-patterns and their detection.
- Magnusson
- 2000
(Show Context)
Citation Context ... within the organisation. Salah, Pauwels, Tavenard & Gevers (2010), reviewed the existing techniques for the discovery of temporal patterns in sensor data and proposed a modified T-Pattern algorithm (=-=Magnusson 2000-=-). This algorithm was tested using the MERL dataset and the presented results outperformed the Lempel-Ziv compression based methods. 3 Employed method In contrast to other mixture of unigrams models t... |

43 | What did you do today? Discovering daily routines from large-scale mobile data,” - Farrahi, Gatica-Perez - 2008 |

43 | Discovering activities to recognize and track in a smart environment,” - Rashidi, Cook, et al. - 2011 |

34 | Discovering Routines from Large-Scale Human Locations using Probabilistic Topic Models.
- Farrahi, Gatica-perez
- 2010
(Show Context)
Citation Context ...h was tested on a short-term scenario, that is, using only seven days of real-world activity data. In contrast, the presented work is tested over long-term sensor data. Farrahi & Gatica-Perez (2008) (=-=Farrahi & Gatica-Perez 2011-=-) used LDA and a modification of the basic LDA model, the Author-Topic Model (Rosen-Zvi, Griffiths, Steyvers & Smyth 2004) to discover daily location-driven routines from a massive dataset of mobile p... |

31 | Image Retrieval on Large-Scale Image Databases - Hörster, Lienhart, et al. - 2007 |

13 | Discovering daily routines from google latitude with topic models”, - Ferrari, Mamei - 2011 |

11 | Identifying rare and subtle behaviours: A weakly supervised joint topic model - Hospedales, Li, et al. - 2011 |

11 | An alignment approach for context prediction tasks in ubicomp environments“. - Sigg, Haseloff, et al. - 2010 |

9 | Relative performance guarantees for approximate inference in latent dirichlet allocation
- Mukherjee, Blei
- 2009
(Show Context)
Citation Context ...ifferences of two common training algorithms. For a good comparison between different inference algorithms that could be used in LDA, we refer the reader to (Asuncion, Welling, Smyth & Teh 2009) and (=-=Mukherjee & Blei 2009-=-). 3.2.1 Variational Bayes (VB) Variational Bayes inference consists on defining a parametric family of distributions that forms a tractable approximation to an intractable true joint distribution. Bl... |

8 | Recovering social networks from massive track datasets. - Connolly, Burns, et al. - 2007 |

8 | Sensor networks for ambient intelligence - Pauwels, Salah, et al. - 2007 |

6 | F.: Supporting activity modeling from activity traces, Expert Systems: The Journal of Knowledge Engineering (submitted) (2010) Intelligent Representation of Research Activities over a Taxonomy 31 - Georgeon, Mille, et al. |

3 | High accuracy context recovery using clustering mechanisms - Phung, Adams, et al. - 2009 |

2 | Modeling and discovering occupancy patterns in sensor networks using latent dirichlet allocation - Castanedo, Aghajan, et al. - 2011 |

2 | Parallelized variational em for latent dirichlet allocation: An experimental evaluation of speed and scalability - Mariote, Medeiros, et al. - 2007 |

1 | Building an occupancy model from sensor networks - Castanedo, López-de-Ipiña, et al. - 2011 |

1 | T-Patterns Revisited: Mining for Temporal - Salah, Pauwels, et al. - 2010 |