{"72907":{"#nid":"72907","#data":{"type":"event","title":"Faculty Candidate Seminar - Multivariate Convex Regression for Value Function Approximation","body":[{"value":"\u003Cp\u003ETITLE: Multivariate Convex Regression for Value Function Approximation\u003C\/p\u003E\u003Cp\u003ESPEAKER: Lauren Hannah\u003C\/p\u003E\u003Cp\u003EABSTRACT:\u003C\/p\u003E\u003Cp\u003EWe propose\n          two new, nonparametric method for multivariate regression\n          subject to convexity or concavity constraints on the response\n          function. \u0026nbsp;Convexity constraints are common in economics,\n          statistics, operations research, financial engineering and\n          optimization, but there is currently no multivariate method\n          that is computationally feasible for more than a few hundred\n          observations. \u0026nbsp;We introduce Convex Adaptive Partitioning (CAP)\n          and Multivariate Bayesian Convex Regression (MBCR), which\n          create a globally convex regression model from locally linear\n          estimates fit on adaptively selected covariate partitions.\u0026nbsp;CAP\n          is computationally efficient, with O(n log(n) log(log(n)))\n          computational complexity, as well as statistically consistent.\n          Although inference for MBCR is more difficult than that of\n          CAP, we show that MBCR is not only consistent but has\n          minimax-optimal adaptive convergence rates. These methods are\n          tested on value function approximation settings in exotic\n          options pricing and response surface methods for simulation\n          optimization.\n    \u003C\/p\u003E\u003Cbr \/\u003E","summary":null,"format":"limited_html"}],"field_subtitle":"","field_summary":"","field_summary_sentence":[{"value":"Multivariate Convex Regression for Value Function Approximation"}],"uid":"27187","created_gmt":"2011-11-22 07:21:48","changed_gmt":"2016-10-08 01:56:41","author":"Anita Race","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2011-11-28T10:00:00-05:00","event_time_end":"2011-11-28T11:00:00-05:00","event_time_end_last":"2011-11-28T11:00:00-05:00","gmt_time_start":"2011-11-28 15:00:00","gmt_time_end":"2011-11-28 16:00:00","gmt_time_end_last":"2011-11-28 16:00:00","rrule":null,"timezone":"America\/New_York"},"extras":[],"groups":[{"id":"1242","name":"School of Industrial and Systems Engineering (ISYE)"}],"categories":[],"keywords":[],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1795","name":"Seminar\/Lecture\/Colloquium"}],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003EJennifer Harris\u003C\/p\u003E","format":"limited_html"}],"email":[],"slides":[],"orientation":[],"userdata":""}}}