Abstract:

Extraordinary amounts of data offer us unprecedented opportunities for scientific discovery and advancement. At the same time, analyzing big data presents unprecedented challenges due to not only volume of data, but also data variety and complexity, as well as the speed with which it must be analyzed. A critical question for the statistics community is how to detect statistical relationship within high volumes of data with complicated structure and turn it into actionable knowledge. Classical statistical models such as linear models or generalized linear models are powerful tools when relationships between the input and output variables are homogeneous, but they can be inadequate when a dataset contains heterogenous patterns, as is often the case with big data. One strategy to address the heterogeneity issue is through the Mixtures-of-Experts (ME) modeling approach. The advantage is that ME is flexible enough to be combined with a variety of different models. Unfortunately, the computing resource requirement is extremely demanding, rendering the computation task formidable for large datasets. In this talk, I will discuss an optimal strategy of subdata selection that preserves maximum information while demanding low computing resource. Under clusterwise linear regression model, it can be shown that the statistical efficiency of the proposed subdata selection algorithm is asymptotically optimal, i.e., there exists no other method with better statistical efficiency than the proposed one when the size of full data is large. Second, as the full data increases in size, the selected subdata preserve the rich information even when the subdata size is fixed.

ZOOM INVITATION

Topic: Colloquium -- Department of Statistics and Actuarial Science -- University of Iowa
Time: Sep 15, 2022 03:15 PM Central Time (US and Canada)

Join Zoom Meeting
https://uiowa.zoom.us/j/91969622615

Meeting ID: 919 6962 2615
One tap mobile
+13017158592,,91969622615# US (Washington DC)
+13092053325,,91969622615# US

Dial by your location
        +1 301 715 8592 US (Washington DC)
        +1 309 205 3325 US
        +1 312 626 6799 US (Chicago)
        +1 646 876 9923 US (New York)
        +1 646 931 3860 US
        +1 669 444 9171 US
        +1 669 900 6833 US (San Jose)
        +1 719 359 4580 US
        +1 253 215 8782 US (Tacoma)
        +1 346 248 7799 US (Houston)
        +1 386 347 5053 US
        +1 564 217 2000 US
Meeting ID: 919 6962 2615
Find your local number: https://uiowa.zoom.us/u/ah4uecMkA

Join by SIP
91969622615@zoomcrc.com

Join by H.323
162.255.37.11 (US West)
162.255.36.11 (US East)
115.114.131.7 (India Mumbai)
115.114.115.7 (India Hyderabad)
213.19.144.110 (Amsterdam Netherlands)
213.244.140.110 (Germany)
103.122.166.55 (Australia Sydney)
103.122.167.55 (Australia Melbourne)
64.211.144.160 (Brazil)
69.174.57.160 (Canada Toronto)
65.39.152.160 (Canada Vancouver)
207.226.132.110 (Japan Tokyo)
149.137.24.110 (Japan Osaka)
Meeting ID: 919 6962 2615