Fictitious El Haddad Datasets

Required Modules

Let us import the modules required to execute B-FADE. The next cell is not necessery if you installed the module via package managers. So you just would skip the cell an jump over.

from sys import path as syspath
from os import path as ospath
syspath.append(ospath.join(ospath.expanduser("~"),
                           '/home/ale/Desktop/b-fade/src'))

import numpy as np
from sklearn.metrics import log_loss
import pandas as pd
from bfade.elhaddad import ElHaddadCurve, ElHaddadBayes
from bfade.dataset import SyntheticDataset

We istantiate a curve whereby the dataset is generated. Then we generate a regular grid and a noisy dataset for later use.

eh = ElHaddadCurve(metrics=np.log10, dk_th=8, ds_w=700, Y=0.9, name="EH-8-700")
eh.inspect(np.linspace(1,1000,1000), scale="log")

14:47:47 - bfade.abstract - WARNING - ElHaddadCurve.inspect
14:47:47 - bfade.util - DEBUG - SHOW PIC: EH-8-700_curve

sd = SyntheticDataset(name="EH-8-700")
sd.make_grid([1,1000], [100,1500], 35, 35, spacing="log")
sd.make_classes(eh)
sd.inspect([1,1000], [100,1500], scale="log", curve=eh, x=np.linspace(1, 1000,1000))

14:47:48 - bfade.dataset - DEBUG - SyntheticDataset.config
14:47:48 - bfade.dataset - DEBUG - SyntheticDataset.make_grid
14:47:48 - bfade.dataset - DEBUG - SyntheticDataset.make_classes
14:47:48 - bfade.dataset - DEBUG - SyntheticDataset.inspect
14:47:48 - bfade.util - DEBUG - SHOW PIC: EH-8-700_data

sd_noise = SyntheticDataset(name="EH-8-700")
sd_noise.make_grid([1,1000], [100,1500], 15, 15, spacing="log")
sd_noise.make_classes(eh)
sd_noise.add_noise(10, 200) # <- make noisy data
sd_noise.crop_points() # <- ensure only physically meaningful data
sd_noise.inspect([1,1000], [100,1500], scale="log", curve=eh, x=np.linspace(1, 1000,1000))

14:47:48 - bfade.dataset - DEBUG - SyntheticDataset.config
14:47:48 - bfade.dataset - DEBUG - SyntheticDataset.make_grid
14:47:48 - bfade.dataset - DEBUG - SyntheticDataset.make_classes
14:47:48 - bfade.dataset - DEBUG - SyntheticDataset.add_noise
14:47:48 - bfade.dataset - DEBUG - SyntheticDataset.crop_points
14:47:48 - bfade.dataset - DEBUG - SyntheticDataset.inspect
14:47:48 - bfade.util - DEBUG - SHOW PIC: EH-8-700_data

Export Data

The operations below exports the datasets as separate .csv files that can be read later.

for d,l in zip([sd, sd_noise], ["EH-7-800", "EH-7-800_Noisy"]):
    pd.DataFrame({"sqrt_area": d.X[:,0],
                  "delta_sigma": d.X[:,1],
                  "failed": d.y,
                  "Y": [0.9]*d.y.shape[0],
                  "test": [0]*d.y.shape[0]}).to_csv("./"+ l +".csv")