Skip to content
stemflow
Using database query to reduce memory load
chenyangkang/stemflow
stemflow
chenyangkang/stemflow
Home
A Brief Introduction
Examples
Examples
A demo of AdaSTEM model
SphereAdaSTEM model: spherical indexing and modeling
Learning curve analysis
Using Maxent as base models
Hurdle model in AdaSTEM
Base model choices
Optimizing stixel size
Lazy loading ensembles
Using database query to reduce memory load
Using database query to reduce memory load
Table of contents
Examples
Download data
Training/test data
Prediction set
Get X and y
First thing first: Spatiotemporal train test split
Initiate AdaSTEM hurdle model
"Traditional" fit with pandas object
Dump data into duckdb database
Using duckdb as input
Using parquet as input
Compare the three method: pd.DataFrame, parquet, and duckdb
On Small dataset
Run test: Training using pd.DataFrame, duckdb, and pandas. Increasing ensemble_fold
Plotting experiment results
Run test: Training using pd.DataFrame, duckdb, and pandas. Increasing n_jobs
Plotting experiment results
On larger dataset
Run test: Training using pd.DataFrame, duckdb, and pandas. Increasing n_jobs
Plotting experiment results
Concluding mark
Tips
Tips
Tips for spatiotemporal indexing
Tips for data types
Tips for different tasks
API Documentation
API Documentation
stemflow.model
stemflow.model
AdaSTEM
SphereAdaSTEM
STEM
static_func_AdaSTEM
Hurdle
dummy_model
stemflow.gridding
stemflow.gridding
Q_blocks
QTree
Sphere_QTree
QuadGrid
stemflow.lazyloading
stemflow.lazyloading
lazyloading
open_db_connection
stemflow.utils
stemflow.utils
jitterrotation
jitterrotation
jitterrotator
sphere
sphere
coordinate_transform
discriminant_formula
distance
Icosahedron
quadtree
sphere_quadtree
plot_gif
generate_random
validation
wrapper
lazyloading
stemflow.model_selection
Fun Visualization
Fun Visualization
Global Bird Diversity
Ruby-crowned_Kinglet
Global NDVI
Global Mean Temperature
Contributing
Code of conduct
Table of contents
Examples
Download data
Training/test data
Prediction set
Get X and y
First thing first: Spatiotemporal train test split
Initiate AdaSTEM hurdle model
"Traditional" fit with pandas object
Dump data into duckdb database
Using duckdb as input
Using parquet as input
Compare the three method: pd.DataFrame, parquet, and duckdb
On Small dataset
Run test: Training using pd.DataFrame, duckdb, and pandas. Increasing ensemble_fold
Plotting experiment results
Run test: Training using pd.DataFrame, duckdb, and pandas. Increasing n_jobs
Plotting experiment results
On larger dataset
Run test: Training using pd.DataFrame, duckdb, and pandas. Increasing n_jobs
Plotting experiment results
Concluding mark