|
Activity Number:
|
466
- Contemporary Statistical Graphics: Methods and Applications
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
Section on Statistical Graphics
|
|
Abstract #313666
|
|
|
Title:
|
A Maximum Entry Based Hypothesis Testing Approach for Estimating Number of Communities
|
|
Author(s):
|
Chetkar Jha* and Ian Barnett and Mingyao Li
|
|
Companies:
|
University of Pennsylvania and University of Pennsylvania and University of Pennsylvania
|
|
Keywords:
|
Stochastic Block Model;
Coherence;
Goodness of Fit;
Higher Criticism;
Number of Communities
|
|
Abstract:
|
Identifying communities in a network is the central problem in network analysis. Community detection algorithms assume the true number of communities to be apriori known. Existing methods for estimating number of communities rely on model selection approaches or spectral norm approaches. We propose a new maximum entry based sequential hypothesis test for estimating the true number of communities. The advantage of our approach is that our method computationally scales better to large datasets compared to most existing approaches. We perform simulations to evaluate and compare our method. We also run our method on single-cell RNA sequencing Data.
|
Authors who are presenting talks have a * after their name.
