Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Derek DeSantis † , Phil Wolfram, Boian Alexandrov Los Alamos National Laboratory, Center for Nonlinear Studies † AMS Annual Meeting, January 2020
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction K¨ oppen-Geiger Model Figure: K¨ oppen-Geiger map of North America (Peel et. al.)
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Problems with K¨ oppen-Geiger Problem Climate depends on more than temperature and precipitation.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Problems with K¨ oppen-Geiger Problem Climate depends on more than temperature and precipitation. Can only resolve land.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Problems with K¨ oppen-Geiger Problem Climate depends on more than temperature and precipitation. Can only resolve land. Does not adapt to changing climate.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Problems with K¨ oppen-Geiger Problem Climate depends on more than temperature and precipitation. Can only resolve land. Does not adapt to changing climate. The cut-offs in model are, to some extent, arbitrary.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Problems with K¨ oppen-Geiger Problem Climate depends on more than temperature and precipitation. Can only resolve land. Does not adapt to changing climate. The cut-offs in model are, to some extent, arbitrary. No universal agreement to how many classes there should be.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Problems with clustering Problem Dependence on algorithm of choice and hyperparameters.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Problems with clustering Problem Dependence on algorithm of choice and hyperparameters. Cluster 1 Consensus Cluster 2 Dataset Clustering Cluster n Figure: Many clusterings combined into a single consensus clustering .
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Problems with clustering Problem Dependence on algorithm of choice and hyperparameters. Cluster 1 Consensus Cluster 2 Dataset Clustering Cluster n Figure: Many clusterings combined into a single consensus clustering . Clustering ill-posed - lack measurement of “trust”.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Problems with clustering Problem Dependence on algorithm of choice and hyperparameters. Cluster 1 Consensus Cluster 2 Dataset Clustering Cluster n Figure: Many clusterings combined into a single consensus clustering . Clustering ill-posed - lack measurement of “trust”. Dependence on “hidden parameters” - scale of data .
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Proposed Solution Solution 1 Leverage discrete wavelet transform to classify across a multitude of scales.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Proposed Solution Solution 1 Leverage discrete wavelet transform to classify across a multitude of scales. 2 Use information theory to discover most important scales to classify on.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Proposed Solution Solution 1 Leverage discrete wavelet transform to classify across a multitude of scales. 2 Use information theory to discover most important scales to classify on. 3 Taking these scales, combine classifications to produce a fuzzy clustering that assess the trust at each point.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Introduction Proposed Solution Solution 1 Leverage discrete wavelet transform to classify across a multitude of scales. 2 Use information theory to discover most important scales to classify on. 3 Taking these scales, combine classifications to produce a fuzzy clustering that assess the trust at each point. CGC 1 CGC 2 Cluster 1 CGC L 1 CGC 1 Consensus Cluster 2 CGC 2 Dataset Clustering CGC L 2 CGC 1 Cluster n CGC 2 CGC L n
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Preliminary Tools Discrete Wavelet Transform and Mutual Information The DWT splits a signal into high and low frequency Low temporal signal captures climatology (seasons, years, decades), DWT while low spatial signal Space captures regional DWT Time features(city, county, DWT state). of Tensor
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Preliminary Tools Discrete Wavelet Transform and Mutual Information The DWT splits a signal into high and low frequency Low temporal signal captures climatology (seasons, years, decades), DWT while low spatial signal Space captures regional DWT Time features(city, county, DWT state). of Tensor Definition Given partitions of data U = { U j } k j =1 , V = { V j } l j =1 , the Mutual Information NI ( U, V ) measures how knowledge of one clustering reduces our uncertainty of the other.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Preliminary Tools L15 Gridded Climate Dataset - Livneh et. al. Gridded climate data set of North America. Grid cell is monthly data from 1950-2013, six kilometers across. Available variables used: precipitation, maximum temperature, minimum temperature.
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) Proposed Solution Solution 1 Leverage discrete wavelet transform to classify across a multitude of scales. 2 Use information theory to discover most important scales to classify on. 3 Taking these scales, combine classifications to produce a fuzzy clustering that assess the trust at each point. CGC 1 CGC 2 Cluster 1 CGC L 1 CGC 1 Consensus Cluster 2 CGC 2 Dataset Clustering CGC L 2 CGC 1 Cluster n CGC 2 CGC L n
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) The Algorithm 1
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) The Algorithm 1 2 DWT DWT DWT
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) The Algorithm 1 2 DWT 3 DWT Stack DWT
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) The Algorithm 1 2 DWT 3 4 DWT Stack Vectorize DWT
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) The Algorithm 1 2 DWT 3 4 5 DWT Stack Vectorize Cluster DWT
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) The Algorithm 1 2 DWT 3 4 5 6 DWT Stack Vectorize Cluster Label DWT
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) Results - Effect of Coarse-Graining Figure: CGC: K-means k = 10, ( ℓ s , ℓ t ) = (1 , 1)
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) Results - Effect of Coarse-Graining Figure: CGC: K-means k = 10, ( ℓ s , ℓ t ) = (4 , 1)
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) Results - Effect of Coarse-Graining Figure: CGC: K-means k = 10, ( ℓ s , ℓ t ) = (1 , 1)
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) Results - Effect of Coarse-Graining Figure: CGC: K-means k = 10, ( ℓ s , ℓ t ) = (1 , 6)
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) Results - Effect of Coarse-Graining Figure: CGC: K-means k = 10, ( ℓ s , ℓ t ) = (1 , 1)
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Coarse-Grain Clustering (CGC) Results - Effect of Coarse-Graining Figure: CGC: K-means k = 10, ( ℓ s , ℓ t ) = (4 , 6)
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Mutual Information Ensemble Reduce (MIER) Proposed Solution Solution 1 Leverage discrete wavelet transform to classify across a multitude of scales. 2 Use information theory to discover most important scales to classify on. 3 Taking these scales, combine classifications to produce a fuzzy clustering that assess the trust at each point. CGC 1 CGC 2 Cluster 1 CGC L 1 CGC 1 Consensus Cluster 2 CGC 2 Dataset Clustering CGC L 2 CGC 1 Cluster n CGC 2 CGC L n
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Mutual Information Ensemble Reduce (MIER) The Algorithm 1
Multiresolution Cluster Analysis—Addressing Trust in Climate Classifications Mutual Information Ensemble Reduce (MIER) The Algorithm 1 2
Recommend
More recommend
