def __init__(self, harmonic: int, detector: str, *args: Any,
**kwargs: Any):
# Base class
super().__init__(*args, **kwargs)
# Configuration
self.harmonic = harmonic
self.main_detector_name = detector
# Properties for determining the event plane resolution
# Here we take all other detectors that aren't the main detector
self.other_detector_names = [
name for name in self.task_config.get("detectors")
if name != self.main_detector_name
]
self.output_ranges = self.task_config.get("output_ranges", None)
# Validation
if self.output_ranges is None:
# Default should be to process each centrality range.
self.output_ranges = [
params.SelectedRange(min, max) for min, max in zip(
np.linspace(0, 100, 11)[:-1],
np.linsapce(0, 100, 11)[1:])
]
# Objects that will be created during the calculation
self.main_detector: Detector
self.other_detectors: List[Detector]
self.resolution: histogram.Histogram1D
self.selected_resolutions: Dict[params.SelectedRange,
Tuple[float, float]] = {}
def lut(val=-1):
"""
Create an empty look up table(LUT)
:param val: If default value is what the lut is initially filled with
if val == 0
the array is all zeros.
if val > 0
the array is set to default value. Clipped to 255.
if val < 0
the array is set to the range [0,255]
if val is a tuple of two values:
we set stretch the range of 0 to 255 to match the range provided.
:return: a LUT.
"""
lut = None
if isinstance(val, list) or isinstance(val, tuple):
start = np.clip(val[0], 0, 255)
stop = np.clip(val[1], 0, 255)
lut = np.around(np.linsapce(start, stop, 256), 0)
lut = np.array(lut, dtype='uint8')
lut = lut.tolist()
elif val == 0:
lut = np.zeros([1, 256]).tolist()[0]
elif val > 0:
val = np.clip(val, 1, 255)
lut = np.ones([1, 256]) * val
lut = np.array(lut, dtype='uint8')
lut = lut.tolist()
elif val < 0:
lut = np.linspace(0, 256, 256)
lut = np.array(lut, dtype='uint8')
lut = lut.tolist()
return lut
def plot_poly_fit():
plot_data()
x=np.linsapce(-60,60,100)
xx = x.reshape(100,1)
XX = np.insert(xx,0,1,axis=1)
xx = poly_feature(xx,power)
xx = feature_normalize(xx,train_mean,train_std)
plt.plot(x,xx@theta_fit,'r--')
def traj(start,end,const):
num_steps = 10
y = np.linspace(start, end, num_steps)
x = np.linsapce(const,const,num_steps)
def gait_trajectory(y, height):
z = np.zeros(num_steps)
for i in range(len(y)):
z[i] = math.sqrt((height)**2 - (y[i])**2)
return z
fig = plt.figure()
ax = fig.gca(projection='3d')
z = gait_trajectory(y, 2)
def optimize(self, f, x0=None, df=None, f_df=None):
if self.num_sample is None:
x = [np.arange(b[0], b[1], self.epsilon) for b in self.bounds]
if self.epsilon is None:
x = [np.linsapce(b[0], b[1], self.num_sample) for b in self.bounds]
X = np.meshgrid(*x)
X = np.array(X).reshape(len(self.bounds), -1).T
f_X = self.cost(X) * f(X)
final_locs = np.argpartition(f_X.reshape(1, len(f_X))[0],
self.save_k)[0:self.save_k]
return np.atleast_2d(X[final_locs]), np.atleast_2d(f_X[final_locs])
def transverse_slice(acc_pos, slice_length, beam, height, nb=10, nh=10):
"""
Returns 3D coordinates of a transverse slice at slice_height. The slice
height is relative to the ship origin but the returned positions are
relative to the accelerometer. To convert back into ship coordinates use
[X]
"""
x = np.linspace(slice_length, slice_length, 1)
y = np.linsapce(-beam / 2, beam / 2, nb)
z = np.linspace(0, height, nh)
x -= acc_pos[0]
y -= acc_pos[1]
z -= acc_pos[2]
XX, YY, ZZ = np.meshgrid(x, y, z)
return XX, YY, ZZ
def prob2
"""
首先录入数据 录入数据并不是矩阵 转换成矩阵
"""
columns = data.shape[1]
X = data.iloc[:, :columns - 1]
y = data.iloc[:, columns - 1:columns]
# 创立一个一维向量 与自变量的取值个数相同
w = pd.Series(np.ones(X.shape[0]))
X.insert(0,'Ones', w)
X = np.matrix(X.values)
y = np.matrix(y.values)
# 创立参数列向量
theta = np.matrix(np.array([0, 0]))
theta, cost, all_theta = gradient_descent(X, y, theta)
x = np.linsapce(data.population.min(),data.population.max(),100) #在最大值和最小值之间均匀的取100个值
f = theta[0,0] + (theta[0,1]*x)#g[0,0]相当于十个100维的列向量 f是最后一次迭代后出来的theta函数
fig,ax = plt.subplots(figsize=(12,8))
#figure使用add_subplot，pyplot使用的是subplot生成一个 figure
#figure里面再包含多个axes，axes本质上数组
#pyplot由两个函数fugure(figsize=指定画板大小)
#和subplot(nrows=,ncols,figsize=(大小))
#fig =plt.figure(figsize=(12,8))
ax.plot(x,f,'r',label='prediction')
#真正画图的数据
ax.scatter(data.population,data.profit,marker='x',label='training data')
#画散点图 原始数据
ax.set_title('predicted profit vs population size')
ax.set_xlabel('population')
ax.set_ylabel('profit')
ax.legend(loc=2)
plt.show()
#画预测出的函数与population之间的关系
def pro3:
"""画出1000次迭代后的tcost值"""
fig,ax = plt.subplot(figsize=(12,8))
ax.plot(np.arrange(1000),cost,'blue',label='error')
#真正画图的数据
ax.set_title('Error vs. Triaining Epoch')
ax.set_xlabel('Iterations')
ax.set_ylabel('Cost')
plt.show()
def pro4:
path1 = os.path.join(r'C:\Users\TenSh\Desktop\MachineLearningEx\ex1\ex1', 'ex1data2.txt')
#os.path.join(os.getcwd(),'data')就是获取当前目录，并组合成新目录
data1 = pd.read_csv(path1, names=['size', 'bedrooms', 'price'])
#均值归一化
data1 = (data1 - data1.mean())/data.std()
# 简洁的语句实现了标准化变换，类似地可以实现任何想要的变换。
# data.mean(axis=0) 输出矩阵为一行,求每列的平均值,同理data.mean(axis=1) 输出矩阵为一列,求每行的平均值
# data.std(axis=0) 输出矩阵为一列,求每列的标准差,同理data.std(axis=1) 输出矩阵为一列,求每行的标准差
# 标准差也成为标准偏差,表示数据的离散程度,和标准差大小成反比
col1 = pd.Series(np.ones(data1.shape[0]))
#对象 不是数组 也不是矩阵？
data1.insert(o,'ones',col1)
cols = data1.shape[1]
X1 = np.matrix(data1.iloc[:,0:cols - 1].values)
#x1 = np.matrix(data2.iloc[:,:])
y1 = np.matrix(data1.iloc[:,cols-1:cols].values)
theta1 = np.matrix([0,0,0])
#构建 自变量 因变量（预测数据） 矩阵
#构建函数矩阵
#得出预测数据 画图
theta1,cost1,all_theta1 = gradient_descent(X1,y1,theta1)
fig1,ax1 = plt.subplots(figsize=(12,8))
ax1.plot(np.arange(1000),cost1,'r',label='multierror')
ax.set_title('multi varible linear regression')
plt.show()
def main():
"""
'''linspace start, stop, numの3つです。 startは数列の開始点を指定する stopは数列の終了点を指定する numは数列の要素の数を指定する ''' x_1 = np.linspace(0, 1, 100) y_1 = x_1 x_2 = np.linspace(0, 1, 100) y_2 = x_2 ** 2 x_3 = np.linspace(0, 1, 100) y_3 = x_3 ** 3 x_4 = np.linspace(0, 1, 100) y_4 = x_4 ** 4 fig = plt.figure() ax_1 = fig.add_subplot(221) #? 2 x 2 so 4 charts first chart ax_2 = fig.add_subplot(222) ax_3 = fig.add_subplot(223) ax_4 = fig.add_subplot(224) ax_1.plot(x_1, y_1) ax_2.plot(x_2, y_2) ax_3.plot(x_3, y_3) ax_4.plot(x_4, y_4) plt.show() print(np.linsapce(0 ,30, 11))
gammas = [0.01, 0.03, 0.05, 0.07, 0.09, 0.12, 0.15] r = -0.1 np.random.seed(0) x = np.r_[np.random.randn(20, 2) - [2, 2], np.random.randn(20, 2) + [2, 2]] y = [0] * 20 + [1] * 20 for gamma in gammas: my_estimator = svm.SVR(kernel = 'sigmoid', gamma = gamma, coef0 = r) my_estimator.fit(x, y) # w = my_estimator.coef_[0] # a = -w[0] / w[1] x_lim = np.linspace(-5, 5) y_lim = np.linsapce(-5, 5) # b = my_estimator.support_vectors_[0] # y_lower = a * x_lim + (b[1] - a * b[0]) b = my_estimator.support_vectors_[-1] y_upper = a * x_lim + (b[1] - a * b[0]) plt.plot(x_lim, y_lim, 'k-*-') plt.plot(x_lim, y_lower, 'bo') plt.plot(x_lim, y_upper, 'r-') plt.scatter(my_estimator.support_vectors_[:, 0], my_estimator.support_vectors_[:, 1], s = 80, facecolors = 'none') plt.scatter(x[:, 0], x[:, 1], c = y, cmap = plt.cm.Paired)
b.date b.data c = dtype([[1,2],[3,4]],dtype=complex) c = np.array([[1,2],[3,4]],dtype=complex) c np.zeros((3,4)) np.zeros([3,4]) np.ones((2,5),dtype=np.int16) np.empty((3,4)) np.arange(0,4,0.5) np.pi pi pi.evalf() pi.evalf(18) pi.evalf(100) np.linsapce(0,2,9) np.linspace(0,2,9) np.linspace(0,2*np.pi,100) x = np.linspace(0,2*np.pi,100) f = np.sin(x) a = np.arange(6) b = np.arange(12).reshape(4,3) b print(b) b = np.arange(12).reshape(2,3,2) b print(np.arange(10000)) a b c = np.linspace(1,6,6) c
def __init__(self, name, Ns, n_class, params = None):
self.name = name
self.Ns = Ns
self.n_class = n_class
if (params is None):
self.params = {}
else:
self.params = params
#For each type of aggregator, set default parameters or check if the provided one are correct.
if (name == 'naive'):
if ("cond_pdf" in self.params):
if (self.params["cond_pdf"].shape[1] != self.n_class):
raise ValueError('Incorrect array size. Conditional probabilities is 3D array with shape (Ns,n_class,n_class)')
else:
self.params["cond_pdf"]=None
if ("prior" in self.params):
if (self.params["prior"].shape != (self.n_class,)):
raise ValueError('Incorrect array size. Prior probabilities is 1D array with shape (n_class,)')
else:
self.params["prior"]=None
elif (name == 'spocc'):
if ("tnorm" in self.params):
if (self.params["tnorm"] not in ['Aczel-Alsina','convex']):
raise ValueError('Unknown tnorm type.')
else:
self.params["tnorm"]='convex'
if ("cond_pdf" in self.params):
if (self.params["cond_pdf"].shape[1] != self.n_class):
raise ValueError('Incorrect array size. Conditional probabilities is 3D array with shape (Ns,n_class,n_class)')
else:
self.params["cond_pdf"]=None
if ("cond_possib" in self.params):
if (self.params["cond_possib"].shape[1] != self.n_class):
raise ValueError('Incorrect array size. Conditional possibilities is 3D array with shape (Ns,n_class,n_class)')
else:
self.params["cond_possib"]=None
if ("hyper" in self.params):
if not(isinstance(self.params["hyper"], Number)):
raise ValueError('Invalid type for parameter hyper which must be a number')
else:
if (self.params["hyper"]<0):
raise ValueError('Invalid value for parameter hyper which must be positive.')
else:
self.params["hyper"] = 0.0
if ("hyper_range" in self.params):
if (self.params["hyper_range"].ndim !=1):
raise ValueError('Invalid array dimension. hyper_range must be a 1D array')
else:
self.params["hyper_range"] = np.logspace(-2,1,101)
elif (name == 'adaspocc'):
self.params["alpha"]=None
self.params["dendro"]=None
if ("clf_scores" in self.params):
if (self.params["clf_scores"].shape != (self.Ns,)):
raise ValueError('Incorrect shape for parameter clf_scores which must have shape (Ns,).')
else:
self.params["clf_scores"] = None
if ("tnorm" in self.params):
if (self.params["tnorm"] not in ['convex','Aczel-Alsina']):
raise ValueError('Unknown tnorm type.')
else:
self.params["tnorm"]='convex'
if ("cond_pdf" in self.params):
if (self.params["cond_pdf"].shape[1] != self.n_class):
raise ValueError('Incorrect array size. Conditional probabilities is 3D array with shape (Ns or more,n_class,n_class)')
else:
self.params["cond_pdf"]=None
if ("cond_possib" in self.params):
if (self.params["cond_possib"].shape[1] != self.n_class):
raise ValueError('Incorrect array size. Conditional possibilities is 3D array with shape (Ns or more,n_class,n_class)')
else:
self.params["cond_possib"]=None
if ("r" in self.params):
if not(isinstance(self.params["r"], Number)):
raise ValueError('Invalid type for parameter r which must be a number')
else:
if (self.params["r"]<0):
raise ValueError('Invalid value for parameter r which must be positive.')
else:
self.params["r"] = 10.0
if ("hyper" in self.params):
if not(isinstance(self.params["hyper"], Number)):
raise ValueError('Invalid type for parameter hyper which must be a number')
else:
if (self.params["hyper"]<0):
raise ValueError('Invalid value for parameter hyper which must be positive.')
else:
self.params["hyper"] = 0.0
if ("hyper_range" in self.params):
if (self.params["hyper_range"].ndim !=1):
raise ValueError('Invalid array dimension. hyper_range must be a 1D array')
else:
self.params["hyper_range"] = np.logspace(-2,1,101)
if ("alpha_range" in self.params):
if (self.params["alpha_range"].ndim !=1):
raise ValueError('Invalid array dimension. alpha_range must be a 1D array')
else:
self.params["alpha_range"] = np.linsapce(0,1,101)
elif (name == 'stacked_logreg'):
if ("hyper" in self.params):
if not(isinstance(self.params["hyper"], Number)):
raise ValueError('Invalid type for parameter hyper which must be a number')
else:
if (self.params["hyper"]<0) :
raise ValueError('Invalid value for parameter hyper which must be positive.')
else:
self.params["hyper"] = 1.0
if ("clf_meta" in self.params):
if not(isinstance(self.params["clf_meta"], LogisticRegression)):
raise ValueError('Invalid type of object. clf_meta must be an instance of LogisticRegression from sklearn.')
else:
self.params["clf_meta"] = LogisticRegression(penalty='l2', C=self.params["hyper"])
if ("regul_range" in self.params):
if (self.params["regul_range"].ndim !=1):
raise ValueError('Invalid array dimension. regul_range must be a 1D array')
else:
self.params["regul_range"] = np.logspace(-2,2,101)
elif (name == 'weighted_vote'):
if ("clf_scores" in self.params):
if (self.params["clf_scores"].shape != (self.Ns,)):
raise ValueError('Incorrect shape for parameter clf_scores which must have shape (Ns,).')
else:
self.params["clf_scores"] = None
if ("r" in self.params):
if not(isinstance(self.params["r"], Number)):
raise ValueError('Invalid type for parameter r which must be a number')
else:
if (self.params["r"]<0):
raise ValueError('Invalid value for parameter r which must be positive.')
else:
self.params["r"] = 10.0
if ("r_range" in self.params):
if (self.params["r_range"].ndim !=1):
raise ValueError('Invalid array dimension. r_range must be a 1D array')
else:
self.params["r_range"] = np.logspace(-1,2,201)
if("expo" in self.params):
if (type(self.params["expo"])!=bool):
raise ValueError('Parameter expo must be a bool.')
else:
self.params["expo"]=True
elif (name == 'selection'):
if ("clf_scores" in self.params):
if (self.params["clf_scores"].shape != (self.Ns,)):
raise ValueError('Incorrect shape for parameter clf_scores which must have shape (Ns,).')
else:
self.params["clf_scores"] = None
if ("select" in self.params):
if (type(self.params["select"]) is not int):
raise ValueError('Invalid type for parameter select which must be an integer.')
else:
self.params["select"] = None
elif (name == 'bayes'):
size = (self.n_class,)
for i in range(Ns):
size += (self.n_class,)
self.params["cond_probas"] = np.zeros(size)
else:
raise ValueError('Unknown type of Aggregator.')
import mplcursors
import Modelo_fast.graficos_matplotly as gm
import numpy as np
lista_arquivos_base = [
"modelo1(42x42)[tipo_1].txt", # 0
"modelo2(42x42)[tipo_1].txt", # 1
"modelo4(42x42)[tipo_1].txt", # 2
"modelo5(42x42)[tipo_1].txt", # 3
"modelo6(42x42)[tipo_1].txt"
] # 4
arquivo_usado = lista_arquivos_base[1]
pesos_dif_orientacao = np.linspace(0, 1, 100)
pesos_distancia = np.linsapce(0, 1, 100)
numero_simulacao = 0
qnt_time_steps_teste = 30
matriz_entropias = []
lista_cores = []
lista_labels = []
for peso_orientacao in pesos_dif_orientacao:
for peso_distancia in pesos_distancia:
numero_simulacao += 1
peso_atual = (peso_orientacao, peso_distancia)
df_resultados = simulacao_com_arquivo(arquivo_usado,