Agmodeling
Statistical modeling tools, to unify model creation and scoring based on python
package agmodeling.setscoring implements a part of the SET method for comparing sensor output as described by :
An Evaluation Tool Kit of Air Quality 1 Micro-Sensing Units (Barak Fishbain1,Uri Lerner, Nuria Castell-Balaguer)
What's New
- (2022/10) change the way to calculate NRMSE
- (2022/09) logging configuration added (v 0.8)
- (2022/09) introducing get_detailed_score() returning all coef (v 0.7) + move to python logger
- (2022/09) correction of warning after pandas version evolution (v 0.6)
- (2019/08) python 3 support (v 0.4)
- (2018/11) First version (v 0.3)
Dependencies
Agmodeling is written to be use with python 2.7 and python 3.6
It requires Pandas, numpy and scipy
It requires Pandas::
pip install pandas
pip install numpy
pip install scipy
Installations
pip install agmodeling
Uses cases
from agmodeling.scoring.set_method import get_IPI_score
import pandas as pd
# logging setup
import logging
consoleHandler = logging.StreamHandler()
logging.basicConfig(
format="%(asctime)s %(levelname)-8s %(message)s",
handlers=[consoleHandler],
level=logging.INFO,
)
file = 'sample_data.xlsx'
print (u'Read excel data file : %s'%file)
df = pd.read_excel(file)
ipi = get_IPI_score(df[u"PM10_REF"], df[u"PM10_MOD_EARTH"])
print (ipi)
Match : RMSE : Pearson : Kendall : Spearman : LFE :: IPI
0.763240 : 0.061937 : 0.909195 : 0.657553 : 0.832455 : 0.990418 :: 0.848801
0.848801
You can run the whole demo inside the package
cd demo
python .\demo_SET_scoring.py
Read excel data file : sample_data.xlsx
containing 2568 data
Score IPI for PM25_RAW
Match : RMSE : Pearson : Kendall : Spearman : LFE :: IPI
0.492835 : 0.869434 : 0.639916 : 0.417968 : 0.575632 : 0.980010 :: 0.539488
Score IPI for PM25_MOD_QUAD
Match : RMSE : Pearson : Kendall : Spearman : LFE :: IPI
0.687539 : 0.374117 : 0.747821 : 0.524258 : 0.695786 : 0.980010 :: 0.710216
Score IPI for PM25_MOD_EARTH
Match : RMSE : Pearson : Kendall : Spearman : LFE :: IPI
0.648910 : 0.337527 : 0.800773 : 0.537126 : 0.713852 : 0.980010 :: 0.723857
Score IPI for PM10_RAW
Match : RMSE : Pearson : Kendall : Spearman : LFE :: IPI
0.486604 : 0.786641 : 0.454199 : 0.269705 : 0.393423 : 0.990388 :: 0.467946
Score IPI for PM10_MOD_QUAD
Match : RMSE : Pearson : Kendall : Spearman : LFE :: IPI
0.742056 : 0.221220 : 0.866073 : 0.612143 : 0.789426 : 0.990388 :: 0.796478
Score IPI for PM10_MOD_EARTH
Match : RMSE : Pearson : Kendall : Spearman : LFE :: IPI
0.763240 : 0.184250 : 0.909195 : 0.657553 : 0.832455 : 0.990388 :: 0.828097
Results :
" RAW MOD_QUAD MOD_EARTH
PM10 0.467946 0.796478 0.828097
PM25 0.539488 0.710216 0.723857"
Fin du programme
