repository for GW potential zonation
Title: A comparative study of machine learning and MCDM Fuzzy-AHP technique to Groundwater potential mapping in the data-scarce region. author: Ranveer Kumar , M.tech(Enginering Geosciences) IIT BHU Varanasi 221005
data guide:
Ahp_accu_check.csv --- contains the true and predicted classes by weighted overlay analysis, for accuracy assesment.
Data_unlabeled.csv ---- data for classification after model training.
Train_Data.csv---------- Data for training Every row represents a features (cell 90m90m). *Columns represents the different parameters as stated below.
1. x = Easting of the centroid of respective cell.
2. y = Northing of the centriod of respective cell.
3.DFR = Distance from River.
4.DFF = Distance from fault.
5.Altitude = Elevation of cells.
6.Geomorph = Geomorphological category.
7.Draw_D = GroundWater Fluctuation (premonsoon-postmonsoon)
8.Drain_D = Drainage Density (Km/sqKm)
9.LS = LS- Factor
10.Litho = Lithology Category.
11.Lineam_D = Lineament density.
12.MRVBF = Multi Resolution index of Valley Bottom Flatness.
13.MRRTF = Multi Resolution Ridge Top Flatness.
14.LULC = Land Use Land Cover Category.
15.SCA = Specific Catchment Area.
16.Profile_C = Profile Curveture
17.Plan_C = Plan Curveture
18.SPI = Specific Precipitation Index
19.Slope = Slope of each feature.
20.TWI = Topographic wetness index.
21.TRI = Topographic Ruggedness index
22.TPI = Topographic position Index.
23.WellYield = wells potential yeilds in Lps. 0 value represent no data.( total 208 well data is available for training and testing)
well_augmentation.py ----- code used for data augmentation to create additional wells in the neighbourhood of existing wells. Train_Data_agm.csv------ Augmented dataset with above mentioned parameters.
python Code for notebooks: (in .ipynb)(raw-- trained with 208 well data, agm -- trained with augmented dataset)
GWP_01_KNN_agm
GWP_01_KNN_raw
GWP_01_RandomForest_agm
GWP_01_RandomForest_raw
GWP01_SVM_agm
GWP01_SVM_raw
data_augmentation.py