diff --git a/Auto_Run.py b/Auto_Run.py
new file mode 100644
index 0000000..f86d14d
--- /dev/null
+++ b/Auto_Run.py
@@ -0,0 +1,148 @@
+import pandas as pd
+import numpy as np
+import yaml
+from sklearn.neighbors import KNeighborsClassifier
+from sklearn.model_selection import train_test_split
+import matplotlib.pyplot as plt
+
+from FS.pso import jfs as jfs_pso
+from FS.ga import jfs as jfs_ga
+from FS.de import jfs as jfs_de
+from FS.ba import jfs as jfs_ba
+from FS.cs import jfs as jfs_cs
+from FS.fa import jfs as jfs_fa
+from FS.fpa import jfs as jfs_fpa
+from FS.sca import jfs as jfs_sca
+from FS.woa import jfs as jfs_woa
+
+
+class DataPipeline(object):
+ """
+ Data Model with Pipeline to Process and Feature Select.
+ """
+ def __init__(self, _ori_data_path:str, _train_test_ratio:float, _str_read_meta_para:str) -> None:
+ """
+ para1:_ori_data_path:Read_RawData_Path(CSV_Format) => Convert to DataFrame
+ para2:_train_test_ratio:Split DataSet Ratio
+ """
+ self._ori_data_path = _ori_data_path
+ self._train_test_ratio = _train_test_ratio
+ self._str_read_meta_para = _str_read_meta_para
+
+ self.data = None
+ self.feat = None
+ self.label = None
+
+ self.fold = None # Contain Splited Train/Test Data
+ self.meta_para = None
+
+ self.sf = None # Select_Feature
+ self.fmdl = None # Feature Model
+
+ def _load_ori_file(self, _ori_path) -> None:
+ '''
+ Load Raw Data File By CSV Path
+ '''
+ self.data = pd.read_csv(_ori_path)
+ self.data = self.data.values
+ #All Feature
+ self.feat = np.asarray(self.data[:, 0:-1])
+ #Predict Y Value
+ self.label = np.asarray(self.data[:, -1])
+
+ def _data_split(self, ratio:float) -> None:
+ '''
+ Setting Train/Test Ratio
+ '''
+ xtrain, xtest, ytrain, ytest = train_test_split(self.feat, self.label, test_size=ratio, stratify=self.label)
+ self.fold = {'xt':xtrain, 'yt':ytrain, 'xv':xtest, 'yv':ytest}
+
+ def _read_algo_parameter(self, _str_read_meta_para:str) ->None:
+ '''
+ Read Meta Para File
+ '''
+ with open('algo_para.yaml', 'r') as _str_read_meta_para:
+ self.meta_para = yaml.full_load(_str_read_meta_para)['meta_para']
+
+ def _build_fmdl(self, algo_name):
+ '''
+ Construct Each Meta Model
+ '''
+ interface_fmdl = {'pso': jfs_pso,'ga':jfs_ga, 'de':jfs_de, 'ba':jfs_ba, 'cs':jfs_cs, 'fa': jfs_fa, 'fpa':jfs_fpa, 'sca':jfs_sca, 'woa':jfs_woa}
+ return interface_fmdl[algo_name]
+
+ def _process_feature_select(self, strMetaName:str, listMetaOpts:list) -> None:
+ """
+ Prepare MetaModel and X Feature Data
+ """
+ # perform feature selection
+ ## Append Fold Data Into OPTS
+ listMetaOpts['fold'] = self.fold
+ #self.fmdl = jfs_ga(self.feat, self.label, listMetaOpts)
+ self.fmdl = self._build_fmdl(strMetaName)(self.feat, self.label, listMetaOpts)
+ self.sf = self.fmdl['sf']
+
+ def _data_feature_select(self, strMetaName:str, listMetaOpts:list)->None:
+ """
+ model with selected features
+ """
+ num_train = np.size(self.fold['xt'], 0)
+ num_valid = np.size(self.fold['xv'], 0)
+ x_train = self.fold['xt'][:, self.sf]
+ y_train = self.fold['yt'].reshape(num_train) # Solve bug
+ x_valid = self.fold['xv'][:, self.sf]
+ y_valid = self.fold['yv'].reshape(num_valid) # Solve bug
+
+ mdl = KNeighborsClassifier(n_neighbors = listMetaOpts['k'])
+ mdl.fit(x_train, y_train)
+
+ # accuracy
+ y_pred = mdl.predict(x_valid)
+ Acc = np.sum(y_valid == y_pred) / num_valid
+ print("Accuracy:", 100 * Acc)
+
+ # number of selected features
+ num_feat = self.fmdl['nf']
+ print("Feature Size:", num_feat)
+
+ def _plot_converege(self, strMetaName:str, listMetaOpts:list)-> None:
+ '''
+ plot convergence
+ '''
+ curve = self.fmdl['c']
+ curve = curve.reshape(np.size(curve,1))
+ x = np.arange(0, listMetaOpts['T'], 1.0) + 1.0
+ fig, ax = plt.subplots()
+ ax.plot(x, curve, 'o-')
+ ax.set_xlabel('Number of Iterations')
+ ax.set_ylabel('Fitness')
+ ax.set_title(strMetaName)
+ ax.grid()
+ plt.show()
+
+ def proceed(self)->None:
+ '''
+ Execute Function
+ '''
+ self._load_ori_file(self._ori_data_path)
+ self._data_split(self._train_test_ratio)
+ self._read_algo_parameter(self._str_read_meta_para)
+
+ for meta in self.meta_para:
+ print(meta['name'])
+ print(meta['opts'])
+ self._process_feature_select(meta['name'], meta['opts'])
+ self._data_feature_select(meta['name'], meta['opts'])
+ self._plot_converege(meta['name'], meta['opts'])
+
+def main():
+ str_read_file_path = './ionosphere.csv'
+ str_read_meta_para = './algo_para.yaml'
+ float_split_ratio = 0.3
+
+ auto_run = DataPipeline(str_read_file_path, float_split_ratio, str_read_meta_para)
+ auto_run.proceed()
+
+if __name__ == '__main__':
+ main()
+
\ No newline at end of file
diff --git a/FS/__basic.py b/FS/__basic.py
new file mode 100644
index 0000000..8512751
--- /dev/null
+++ b/FS/__basic.py
@@ -0,0 +1,33 @@
+import numpy as np
+from numpy.random import rand
+from FS.__tran_func import tran_func
+
+def init_position(lb, ub, N, dim):
+ X = np.zeros([N, dim], dtype='float')
+ for i in range(N):
+ for d in range(dim):
+ X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
+
+ return X
+
+def binary_conversion(X, thres, N, dim, trans_func=tran_func.sl_trans):
+ Xbin = np.zeros([N, dim], dtype='int')
+ for i in range(N):
+ for d in range(dim):
+ X_trans = trans_func(X[i,d])
+ if X_trans > thres:
+ Xbin[i,d] = 1
+ else:
+ Xbin[i,d] = 0
+
+ return Xbin
+
+
+def boundary(x, lb, ub):
+ if x < lb:
+ x = lb
+ if x > ub:
+ x = ub
+
+ return x
+
diff --git a/FS/__tran_func.py b/FS/__tran_func.py
new file mode 100644
index 0000000..af88323
--- /dev/null
+++ b/FS/__tran_func.py
@@ -0,0 +1,10 @@
+import numpy as np
+class tran_func():
+
+ @staticmethod
+ def l_trans(val):
+ return val
+
+ @staticmethod
+ def sl_trans(val):
+ return 1 / (1+np.exp(-2 * val))
\ No newline at end of file
diff --git a/FS/ba.py b/FS/ba.py
index 1ab3e80..41e3c96 100644
--- a/FS/ba.py
+++ b/FS/ba.py
@@ -3,42 +3,13 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
-
-
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
-
+from FS.__basic import init_position, binary_conversion, boundary
+
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
fmax = 2 # maximum frequency
fmin = 0 # minimum frequency
diff --git a/FS/cs.py b/FS/cs.py
index da6998d..1001bd6 100644
--- a/FS/cs.py
+++ b/FS/cs.py
@@ -3,39 +3,10 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
+from FS.__basic import init_position, binary_conversion, boundary
import math
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
-
-
# Levy Flight
def levy_distribution(beta, dim):
# Sigma
@@ -54,8 +25,8 @@ def levy_distribution(beta, dim):
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
Pa = 0.25 # discovery rate
alpha = 1 # constant
diff --git a/FS/de.py b/FS/de.py
index 9903258..77019b2 100644
--- a/FS/de.py
+++ b/FS/de.py
@@ -3,42 +3,13 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
-
-
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
+from FS.__basic import init_position, binary_conversion, boundary
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
CR = 0.9 # crossover rate
F = 0.5 # factor
diff --git a/FS/fa.py b/FS/fa.py
index 85cb0bf..792c390 100644
--- a/FS/fa.py
+++ b/FS/fa.py
@@ -3,42 +3,13 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
-
-
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
+from FS.__basic import init_position, binary_conversion, boundary
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
alpha = 1 # constant
beta0 = 1 # light amplitude
diff --git a/FS/fpa.py b/FS/fpa.py
index f21a6ec..7627c60 100644
--- a/FS/fpa.py
+++ b/FS/fpa.py
@@ -3,37 +3,11 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
-import math
-
-
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
+from FS.__basic import init_position, binary_conversion, boundary
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
+import math
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
# Levy Flight
@@ -54,8 +28,8 @@ def levy_distribution(beta, dim):
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
beta = 1.5 # levy component
P = 0.8 # switch probability
diff --git a/FS/ga.py b/FS/ga.py
index ed48729..2ba48a9 100644
--- a/FS/ga.py
+++ b/FS/ga.py
@@ -1,28 +1,7 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
-
-
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
+from FS.__basic import init_position, binary_conversion, boundary
def roulette_wheel(prob):
num = len(prob)
@@ -38,8 +17,8 @@ def roulette_wheel(prob):
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
CR = 0.8 # crossover rate
MR = 0.01 # mutation rate
diff --git a/FS/gwo.py b/FS/gwo.py
index 0bdb35c..7244fb3 100644
--- a/FS/gwo.py
+++ b/FS/gwo.py
@@ -3,42 +3,13 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
+from FS.__basic import init_position, binary_conversion, boundary
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
-
-
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
N = opts['N']
diff --git a/FS/hho.py b/FS/hho.py
index e54c64c..ec49e4f 100644
--- a/FS/hho.py
+++ b/FS/hho.py
@@ -3,39 +3,10 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
+from FS.__basic import init_position, binary_conversion, boundary
import math
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
-
-
def levy_distribution(beta, dim):
# Sigma
nume = math.gamma(1 + beta) * np.sin(np.pi * beta / 2)
@@ -53,8 +24,8 @@ def levy_distribution(beta, dim):
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
beta = 1.5 # levy component
diff --git a/FS/ja.py b/FS/ja.py
index 3592a10..48bdd1f 100644
--- a/FS/ja.py
+++ b/FS/ja.py
@@ -3,42 +3,14 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
+from FS.__basic import init_position, binary_conversion, boundary
-
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
N = opts['N']
diff --git a/FS/pso.py b/FS/pso.py
index 495e11d..7491c9d 100644
--- a/FS/pso.py
+++ b/FS/pso.py
@@ -1,15 +1,7 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
-
-
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
+from FS.__basic import init_position, binary_conversion, boundary
def init_velocity(lb, ub, N, dim):
@@ -26,33 +18,12 @@ def init_velocity(lb, ub, N, dim):
V[i,d] = Vmin[0,d] + (Vmax[0,d] - Vmin[0,d]) * rand()
return V, Vmax, Vmin
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
w = 0.9 # inertia weight
c1 = 2 # acceleration factor
diff --git a/FS/sca.py b/FS/sca.py
index 9e1616e..6c3418e 100644
--- a/FS/sca.py
+++ b/FS/sca.py
@@ -3,42 +3,13 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
-
-
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
+from FS.__basic import init_position, binary_conversion, boundary
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
alpha = 2 # constant
diff --git a/FS/ssa.py b/FS/ssa.py
index 53f8cce..df5841c 100644
--- a/FS/ssa.py
+++ b/FS/ssa.py
@@ -3,42 +3,14 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
+from FS.__basic import init_position, binary_conversion, boundary
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
-
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
N = opts['N']
diff --git a/FS/woa.py b/FS/woa.py
index a6f7a8f..403324e 100644
--- a/FS/woa.py
+++ b/FS/woa.py
@@ -3,42 +3,12 @@
import numpy as np
from numpy.random import rand
from FS.functionHO import Fun
-
-
-def init_position(lb, ub, N, dim):
- X = np.zeros([N, dim], dtype='float')
- for i in range(N):
- for d in range(dim):
- X[i,d] = lb[0,d] + (ub[0,d] - lb[0,d]) * rand()
-
- return X
-
-
-def binary_conversion(X, thres, N, dim):
- Xbin = np.zeros([N, dim], dtype='int')
- for i in range(N):
- for d in range(dim):
- if X[i,d] > thres:
- Xbin[i,d] = 1
- else:
- Xbin[i,d] = 0
-
- return Xbin
-
-
-def boundary(x, lb, ub):
- if x < lb:
- x = lb
- if x > ub:
- x = ub
-
- return x
-
+from FS.__basic import init_position, binary_conversion, boundary
def jfs(xtrain, ytrain, opts):
# Parameters
- ub = 1
- lb = 0
+ ub = opts['ub']
+ lb = opts['lb']
thres = 0.5
b = 1 # constant
diff --git a/README.md b/README.md
index b5c363a..a8df18d 100644
--- a/README.md
+++ b/README.md
@@ -215,4 +215,7 @@ plt.show()
| 01 | `ga` | [Genetic Algorithm](/Description.md#genetic-algorithm-ga) | - | Yes |
-
+## Auto run
+* You can run multiple-model in one python code(Auto_Run.py), compare with model result.
+* Model List: pso, ga, de, ba, cs, fa, fpa, sca, woa
+* Setting File: algo_para.yaml
diff --git a/algo_para.yaml b/algo_para.yaml
new file mode 100644
index 0000000..dfe7c12
--- /dev/null
+++ b/algo_para.yaml
@@ -0,0 +1,43 @@
+meta_para: ## Ref: https://github.com/tongysmember/Wrapper-Feature-Selection-Toolbox-Python/blob/main/Description.md#differential-evolution-de
+- name: pso
+ # K : k-value in KNN
+ # N : number of particles
+ # T : maximum number of iterations
+ opts: {'k': 5, 'N': 10, 'T': 10, 'ub': 1, 'lb': 0}
+- name: ga
+ # K : k-value in KNN
+ # N : number of particles
+ # T : maximum number of iterations
+ # T
+ # MR
+ opts: {'k': 5, 'N': 10, 'T': 10, 'CR': 0.8, 'MR': 0.01, 'ub': 1, 'lb': 0}
+- name: de
+ # CR : crossover rate
+ # F : constant factor
+ opts: {'k': 5, 'N': 10, 'T': 10, 'CR': 0.9, 'F': 0.5, 'ub': 1, 'lb': 0}
+- name: ba
+ #fmax = 2 # maximum frequency
+ #fmin = 0 # minimum frequency
+ #alpha = 0.9 # constant
+ #gamma = 0.9 # constant
+ #A = 2 # maximum loudness
+ #r = 1 # maximum pulse rate
+ opts: {'k': 5, 'N': 10, 'T': 10, 'fmax': 2, 'fmin': 0, 'alpha': 0.9, 'gamma': 0.9, 'A': 2, 'r': 1, 'ub': 1, 'lb': 0}
+- name: cs
+ #Pa = 0.25 # discovery rate
+ opts: {'k': 5, 'N': 10, 'T': 10, 'Pa': 0.25, 'ub': 1, 'lb': 0}
+- name: fa
+ #alpha = 1 # constant
+ #beta0 = 1 # light amplitude
+ #gamma = 1 # absorbtion coefficient
+ #theta = 0.97 # control alpha
+ opts: {'k': 5, 'N': 10, 'T': 10, 'alpha': 1, 'beta0': 1, 'gamma': 1, 'theta': 0.97, 'ub': 1, 'lb': 0}
+- name: fpa
+ #P = 0.8 # switch probability
+ opts: {'k': 5, 'N': 10, 'T': 10, 'P': 0.8, 'ub': 1, 'lb': 0}
+- name: sca
+ #alpha = 2 # constant
+ opts: {'k': 5, 'N': 10, 'T': 10, 'alpha': 2, 'ub': 1, 'lb': 0}
+- name: woa
+ #b = 1 # constant
+ opts: {'k': 5, 'N': 10, 'T': 10, 'b': 1, 'ub': 1, 'lb': 0}
\ No newline at end of file