Latent Self-Exciting Point Process Model for Spatial-Temporal Networks

Latent Self-Exciting Point Process Model for Spatial-Temporal Networks. Cho, Y., Galstyan, A., Brantingham, P. J., & Tita, G. Discrete and Continuous Dynamical Systems - Series B, 19(5):1335–1354, April, 2014. arXiv: 1302.2671

Paper doi abstract bibtex

We propose a latent self-exciting point process model that describes geographically distributed interactions between pairs of entities. In contrast to most existing approaches that assume fully observable interactions, here we consider a scenario where certain interaction events lack information about participants. Instead, this information needs to be inferred from the available observations. We develop an efficient approximate algorithm based on variational expectation-maximization to infer unknown participants in an event given the location and the time of the event. We validate the model on synthetic as well as real-world data, and obtain very promising results on the identity-inference task. We also use our model to predict the timing and participants of future events, and demonstrate that it compares favorably with baseline approaches.

@article{cho_latent_2014,
	title = {Latent {Self}-{Exciting} {Point} {Process} {Model} for {Spatial}-{Temporal} {Networks}},
	volume = {19},
	issn = {1531-3492},
	url = {http://arxiv.org/abs/1302.2671},
	doi = {10.3934/dcdsb.2014.19.1335},
	abstract = {We propose a latent self-exciting point process model that describes geographically distributed interactions between pairs of entities. In contrast to most existing approaches that assume fully observable interactions, here we consider a scenario where certain interaction events lack information about participants. Instead, this information needs to be inferred from the available observations. We develop an efficient approximate algorithm based on variational expectation-maximization to infer unknown participants in an event given the location and the time of the event. We validate the model on synthetic as well as real-world data, and obtain very promising results on the identity-inference task. We also use our model to predict the timing and participants of future events, and demonstrate that it compares favorably with baseline approaches.},
	number = {5},
	urldate = {2017-01-16},
	journal = {Discrete and Continuous Dynamical Systems - Series B},
	author = {Cho, Yoon-Sik and Galstyan, Aram and Brantingham, P. Jeffrey and Tita, George},
	month = apr,
	year = {2014},
	note = {arXiv: 1302.2671},
	pages = {1335--1354},
}

Downloads: 0

{"_id":"jHQMYkXQNt9jCebCH","bibbaseid":"cho-galstyan-brantingham-tita-latentselfexcitingpointprocessmodelforspatialtemporalnetworks-2014","downloads":0,"creationDate":"2015-08-10T01:04:25.558Z","title":"Latent Self-Exciting Point Process Model for Spatial-Temporal Networks","author_short":["Cho, Y.","Galstyan, A.","Brantingham, P. J.","Tita, G."],"year":2014,"bibtype":"article","biburl":"https://bibbase.org/zotero/wybert","bibdata":{"bibtype":"article","type":"article","title":"Latent Self-Exciting Point Process Model for Spatial-Temporal Networks","volume":"19","issn":"1531-3492","url":"http://arxiv.org/abs/1302.2671","doi":"10.3934/dcdsb.2014.19.1335","abstract":"We propose a latent self-exciting point process model that describes geographically distributed interactions between pairs of entities. In contrast to most existing approaches that assume fully observable interactions, here we consider a scenario where certain interaction events lack information about participants. Instead, this information needs to be inferred from the available observations. We develop an efficient approximate algorithm based on variational expectation-maximization to infer unknown participants in an event given the location and the time of the event. We validate the model on synthetic as well as real-world data, and obtain very promising results on the identity-inference task. We also use our model to predict the timing and participants of future events, and demonstrate that it compares favorably with baseline approaches.","number":"5","urldate":"2017-01-16","journal":"Discrete and Continuous Dynamical Systems - Series B","author":[{"propositions":[],"lastnames":["Cho"],"firstnames":["Yoon-Sik"],"suffixes":[]},{"propositions":[],"lastnames":["Galstyan"],"firstnames":["Aram"],"suffixes":[]},{"propositions":[],"lastnames":["Brantingham"],"firstnames":["P.","Jeffrey"],"suffixes":[]},{"propositions":[],"lastnames":["Tita"],"firstnames":["George"],"suffixes":[]}],"month":"April","year":"2014","note":"arXiv: 1302.2671","pages":"1335–1354","bibtex":"@article{cho_latent_2014,\n\ttitle = {Latent {Self}-{Exciting} {Point} {Process} {Model} for {Spatial}-{Temporal} {Networks}},\n\tvolume = {19},\n\tissn = {1531-3492},\n\turl = {http://arxiv.org/abs/1302.2671},\n\tdoi = {10.3934/dcdsb.2014.19.1335},\n\tabstract = {We propose a latent self-exciting point process model that describes geographically distributed interactions between pairs of entities. In contrast to most existing approaches that assume fully observable interactions, here we consider a scenario where certain interaction events lack information about participants. Instead, this information needs to be inferred from the available observations. We develop an efficient approximate algorithm based on variational expectation-maximization to infer unknown participants in an event given the location and the time of the event. We validate the model on synthetic as well as real-world data, and obtain very promising results on the identity-inference task. We also use our model to predict the timing and participants of future events, and demonstrate that it compares favorably with baseline approaches.},\n\tnumber = {5},\n\turldate = {2017-01-16},\n\tjournal = {Discrete and Continuous Dynamical Systems - Series B},\n\tauthor = {Cho, Yoon-Sik and Galstyan, Aram and Brantingham, P. Jeffrey and Tita, George},\n\tmonth = apr,\n\tyear = {2014},\n\tnote = {arXiv: 1302.2671},\n\tpages = {1335--1354},\n}\n\n","author_short":["Cho, Y.","Galstyan, A.","Brantingham, P. J.","Tita, G."],"key":"cho_latent_2014","id":"cho_latent_2014","bibbaseid":"cho-galstyan-brantingham-tita-latentselfexcitingpointprocessmodelforspatialtemporalnetworks-2014","role":"author","urls":{"Paper":"http://arxiv.org/abs/1302.2671"},"metadata":{"authorlinks":{"galstyan, a":"https://www.isi.edu/results/publications/"}},"html":""},"search_terms":["latent","self","exciting","point","process","model","spatial","temporal","networks","cho","galstyan","brantingham","tita"],"keywords":[],"authorIDs":["545717002abc8e9f3700004b","5e3905badc5b8ade010000d5","8Z7PHZTPG7ZSJoadv","AuBbkw5Pa4SHKYbeN","tWgn5aMc7nNcDu9n6"],"dataSources":["dDcybAPCHBAZER5nc","NEs7rST9He2dCAf9e","AfBybHaxyt33K5MBP","TJkbwzD8s2wCxBy6Y"]}