Pinned Repositories
accuracy-sensitivity-specificity-precision
Unbalanced targets require special attention to the sensitivity vs specificity tradeoff, and the sensitivity vs precision tradeoff. Here I use two examples, cancer and loan defaults, to demonstrate these tradeoffs in R.
ancient_amazonia_archaeology
machine learning model to find geoglyphs and Amazonian Dark Earth sites in Amazonia
arrow_vs_dart
Classifier for North American arrows versus darts
baseball-heat-map
Use baseball data from Mike Trout to model exit velocity as a function of pitch location. Visualize with a heat map in R
cross-validation-glm-vs-lmer-vs-brms
Repeated k-fold cross validation to evaluate glm (linear) versus lmer (mixed) versus brms (bayesian) models
end2end-all-conv
End-to-end training for breast cancer diagnosis using deep all convolutional networks
feature-selection-4ways
Do variable selection in 4 different ways: 1) rank models by AICc, 2) leave one out or k-fold cross validation, 3) Bayesian variable selection, and 4) use caret's recursive feature elimination.
keras-logistic-regression
Show how a Keras neural network simplifies to a logistic regression
neural-network-regularization
Use Keras in R to purposefully overfit the MNIST dataset. Then improve generalizability with dropout and regularization
shiny-hunter-gatherer-area-scaling
Use R shiny app that uses latitude and population to predict home range area of hunter-gatherers (Binford dataset)
RobertSWalker's Repositories
RobertSWalker/cross-validation-glm-vs-lmer-vs-brms
Repeated k-fold cross validation to evaluate glm (linear) versus lmer (mixed) versus brms (bayesian) models
RobertSWalker/accuracy-sensitivity-specificity-precision
Unbalanced targets require special attention to the sensitivity vs specificity tradeoff, and the sensitivity vs precision tradeoff. Here I use two examples, cancer and loan defaults, to demonstrate these tradeoffs in R.
RobertSWalker/baseball-heat-map
Use baseball data from Mike Trout to model exit velocity as a function of pitch location. Visualize with a heat map in R
RobertSWalker/end2end-all-conv
End-to-end training for breast cancer diagnosis using deep all convolutional networks
RobertSWalker/feature-selection-4ways
Do variable selection in 4 different ways: 1) rank models by AICc, 2) leave one out or k-fold cross validation, 3) Bayesian variable selection, and 4) use caret's recursive feature elimination.
RobertSWalker/keras-logistic-regression
Show how a Keras neural network simplifies to a logistic regression
RobertSWalker/neural-network-regularization
Use Keras in R to purposefully overfit the MNIST dataset. Then improve generalizability with dropout and regularization
RobertSWalker/shiny-hunter-gatherer-area-scaling
Use R shiny app that uses latitude and population to predict home range area of hunter-gatherers (Binford dataset)
RobertSWalker/ancient_amazonia_archaeology
machine learning model to find geoglyphs and Amazonian Dark Earth sites in Amazonia
RobertSWalker/arrow_vs_dart
Classifier for North American arrows versus darts
RobertSWalker/bncompare
Bayesian networks and comparative causal models of human complexity across anthropological datasets
RobertSWalker/dplace-data
The data repository for the D-PLACE Project (Database of Places, Language, Culture and Environment)
RobertSWalker/elv
Data and code for extralethal violence project
RobertSWalker/gg3d
3D modeling and estimates of earthworking for Amazonian geoglyphs
RobertSWalker/hgworld
Reconstructing hunter-gatherer population densities
RobertSWalker/Python
School projects in Python that use the arcade library (http://arcade.academy/)
RobertSWalker/terra2
Modeling indigenous homicides in Brazil
RobertSWalker/terras
Modeling the social drivers of environmental sustainability among Amazonian indigenous lands using Bayesian networks
RobertSWalker/uncontacted
Growth of isolated indigenous metapopulations
RobertSWalker/wnai_bayesnet
Bayesian network and causal inferences of complexity in the Western North American Indian dataset