def compute_mse(regressor, horizon):
# get wind park and corresponding target.
windpark = NREL().get_windpark(NREL.park_id['tehachapi'], 3, 2004, 2005)
target = windpark.get_target()
# use power mapping for pattern-label mapping.
feature_window = 3
mapping = PowerMapping()
X = mapping.get_features_park(windpark, feature_window, horizon)
y = mapping.get_labels_turbine(target, feature_window, horizon)
# train roughly for the year 2004, test for 2005.
train_to = int(math.floor(len(X) * 0.5))
test_to = len(X)
train_step, test_step = 25, 25
X_train=X[:train_to:train_step]
y_train=y[:train_to:train_step]
X_test=X[train_to:test_to:test_step]
y_test=y[train_to:test_to:test_step]
if(regressor == 'svr'):
reg = SVR(kernel='rbf', epsilon=0.1, C = 100.0,\
gamma = 0.0001).fit(X_train,y_train)
mse = mean_squared_error(reg.predict(X_test),y_test)
elif(regressor == 'knn'):
reg = KNeighborsRegressor(10, 'uniform').fit(X_train,y_train)
mse = mean_squared_error(reg.predict(X_test),y_test)
return mse
def compute_mse(regressor, param):
# get wind park and corresponding target. forecast is for the target
# turbine
park_id = NREL.park_id['tehachapi']
windpark = NREL().get_windpark(park_id, 3, 2004)
target = windpark.get_target()
# use power mapping for pattern-label mapping. Feature window length
# is 3 time steps and time horizon (forecast) is 3 time steps.
feature_window = 6
horizon = 3
mapping = PowerMapping()
X = mapping.get_features_park(windpark, feature_window, horizon)
Y = mapping.get_labels_turbine(target, feature_window, horizon)
# train roughly for the year 2004.
train_to = int(math.floor(len(X) * 0.5))
# test roughly for the year 2005.
test_to = len(X)
# train and test only every fifth pattern, for performance.
train_step, test_step = 5, 5
if(regressor == 'rf'):
# random forest regressor
reg = RandomForestRegressor(n_estimators=param, criterion='mse')
reg = reg.fit(X[0:train_to:train_step], Y[0:train_to:train_step])
y_hat = reg.predict(X[train_to:test_to:test_step])
elif(regressor == 'knn'):
# TODO the regressor does not need to be newly trained in
# the case of KNN
reg = KNeighborsRegressor(param, 'uniform')
# fitting the pattern-label pairs
reg = reg.fit(X[0:train_to:train_step], Y[0:train_to:train_step])
y_hat = reg.predict(X[train_to:test_to:test_step])
else:
raise Exception("No regressor set.")
# naive is also known as persistence model.
naive_hat = zeros(len(y_hat), dtype = float32)
for i in range(0, len(y_hat)):
# naive label is the label as horizon time steps before.
# we have to consider to use only the fifth label here, too.
naive_hat[i] = Y[train_to + (i * test_step) - horizon]
# computing the mean squared errors of Linear and naive prediction.
mse_y_hat, mse_naive_hat = 0, 0
for i in range(0, len(y_hat)):
y = Y[train_to + (i * test_step)]
mse_y_hat += (y_hat[i] - y) ** 2
mse_naive_hat += (naive_hat[i] - y) ** 2
mse_y_hat /= float(len(y_hat))
mse_naive_hat /= float(len(y_hat))
return mse_y_hat, mse_naive_hat
def compute_mse(regressor, param):
# get wind park and corresponding target. forecast is for the target
# turbine
park_id = NREL.park_id['tehachapi']
windpark = NREL().get_windpark(park_id, 3, 2004)
target = windpark.get_target()
# use power mapping for pattern-label mapping. Feature window length
# is 3 time steps and time horizon (forecast) is 3 time steps.
feature_window = 6
horizon = 3
mapping = PowerMapping()
X = mapping.get_features_park(windpark, feature_window, horizon)
Y = mapping.get_labels_turbine(target, feature_window, horizon)
# train roughly for the year 2004.
train_to = int(math.floor(len(X) * 0.5))
# test roughly for the year 2005.
test_to = len(X)
# train and test only every fifth pattern, for performance.
train_step, test_step = 5, 5
if (regressor == 'rf'):
# random forest regressor
reg = RandomForestRegressor(n_estimators=param, criterion='mse')
reg = reg.fit(X[0:train_to:train_step], Y[0:train_to:train_step])
y_hat = reg.predict(X[train_to:test_to:test_step])
elif (regressor == 'knn'):
# TODO the regressor does not need to be newly trained in
# the case of KNN
reg = KNeighborsRegressor(param, 'uniform')
# fitting the pattern-label pairs
reg = reg.fit(X[0:train_to:train_step], Y[0:train_to:train_step])
y_hat = reg.predict(X[train_to:test_to:test_step])
else:
raise Exception("No regressor set.")
# naive is also known as persistence model.
naive_hat = zeros(len(y_hat), dtype=float32)
for i in range(0, len(y_hat)):
# naive label is the label as horizon time steps before.
# we have to consider to use only the fifth label here, too.
naive_hat[i] = Y[train_to + (i * test_step) - horizon]
# computing the mean squared errors of Linear and naive prediction.
mse_y_hat, mse_naive_hat = 0, 0
for i in range(0, len(y_hat)):
y = Y[train_to + (i * test_step)]
mse_y_hat += (y_hat[i] - y)**2
mse_naive_hat += (naive_hat[i] - y)**2
mse_y_hat /= float(len(y_hat))
mse_naive_hat /= float(len(y_hat))
return mse_y_hat, mse_naive_hat
def experiment(method, windpark, windpark_test, damaged, rate):
args, nseries = argfuncs[method](windpark)
reconstructed = interpolate(damaged, **args)
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
turbines = windpark.get_turbines()
for t in range(len(turbines)):
turbines[t].add_measurements(\
repair_nrel(turbines[t].get_measurements()[:10000]))
# this is the target turbine, use the reconstructed here.
turbines[-1].add_measurements(reconstructed)
feature_window, horizon = 3,3
mapping = PowerMapping()
# with damaged
X = mapping.get_features_park(windpark, feature_window, horizon)
Y = mapping.get_labels_turbine(target, feature_window, horizon)
train_to = int(math.floor(len(X)))
train_step, test_step = 1, 1
reg = linear_model.LinearRegression()
reg = reg.fit(X[0:train_to:train_step], Y[0:train_to:train_step])
# USE THE 2005 YEAR FOR TESTING, WITHOUT DAMAGE
# predict on second year without damage
turbines = windpark_test.get_turbines()
for t in turbines:
t.add_measurements(repair_nrel(t.get_measurements()[:10000]))
target_test = windpark_test.get_target()
XT = mapping.get_features_park(windpark_test, feature_window, horizon)
test_to = int(math.floor(len(XT)))
YT = mapping.get_labels_turbine(target_test, feature_window, horizon)[:test_to]
y_hat = reg.predict(XT[:test_to])
mse_y_hat = 0
for i in range(0, len(y_hat)):
y = YT[i]
mse_y_hat += (y_hat[i] - y) ** 2
mse_y_hat /= float(len(y_hat))
return mse_y_hat
def setUpClass(cls):
ds = NREL()
cls.turbine = ds.get_turbine(NREL.park_id['tehachapi'], 2004, 2005)
cls.windpark = ds.get_windpark(NREL.park_id['tehachapi'], 3, 2004,
2005)
cls.pmapping = PowerMapping()
cls.pdmapping = PowerDiffMapping()
def fun(citynum,methodnum,K):
park_id = NREL.park_id[cityname[citynum]]
windpark = NREL().get_windpark(park_id, 10, 2004,2006)
pla=[]
kk=windpark.get_turbines()
for i in range(len(kk)):
pla.append(kk[i].idx)
feature_window, horizon = 3, 3
mapping = PowerMapping()
data_1 = np.array(mapping.get_features_park(windpark, feature_window, horizon))
data_train = np.array(mapping.get_features_park(windpark, 1, 1))
lendata=len(data_1)
data1 = data_1[:lendata:3]
l1=len(data_train)
data_train1=data_train[:l1:3]
half=int(math.floor(len(data1) * 0.5))
traindata_1=data_train1[0:half,:]
traindata1=np.transpose(traindata_1)
traindata1=preprocessing.scale(np.array(traindata1),with_mean=True,with_std=True)
if methodnum==0:
ans = KMeans(n_clusters=K, random_state=0).fit(traindata1).predict(traindata1)
if methodnum==1:
ans = SpectralClustering(n_clusters=K, random_state=0).fit_predict(traindata1)
if methodnum==2:
ans = AgglomerativeClustering(n_clusters=K).fit_predict(traindata1)
if methodnum==3:
ans = Birch(n_clusters=K).fit_predict(traindata1)
if methodnum==4:
ans = DBSCAN(eps = 0.1).fit_predict(traindata1)
fo = open('cluster10/'+cityname[citynum]+method[methodnum]+str(K)+'.csv','w', newline='')
csv_write = csv.writer(fo,dialect='excel')
for i in range(len(ans)):
cc=[];
cc.append(pla[i])
cc.append(ans[i])
csv_write.writerow(cc)
fo.close()
def experiment(method, windpark, windpark_test, damaged, rate):
args, nseries = argfuncs[method](windpark)
reconstructed = interpolate(damaged, **args)
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
turbines = windpark.get_turbines()
for t in range(len(turbines)):
turbines[t].add_measurements(\
repair_nrel(turbines[t].get_measurements()[:10000]))
# this is the target turbine, use the reconstructed here.
turbines[-1].add_measurements(reconstructed)
feature_window, horizon = 3, 3
mapping = PowerMapping()
# with damaged
X = mapping.get_features_park(windpark, feature_window, horizon)
Y = mapping.get_labels_turbine(target, feature_window, horizon)
train_to = int(math.floor(len(X)))
train_step, test_step = 1, 1
reg = linear_model.LinearRegression()
reg = reg.fit(X[0:train_to:train_step], Y[0:train_to:train_step])
# USE THE 2005 YEAR FOR TESTING, WITHOUT DAMAGE
# predict on second year without damage
turbines = windpark_test.get_turbines()
for t in turbines:
t.add_measurements(repair_nrel(t.get_measurements()[:10000]))
target_test = windpark_test.get_target()
XT = mapping.get_features_park(windpark_test, feature_window, horizon)
test_to = int(math.floor(len(XT)))
YT = mapping.get_labels_turbine(target_test, feature_window,
horizon)[:test_to]
y_hat = reg.predict(XT[:test_to])
mse_y_hat = 0
for i in range(0, len(y_hat)):
y = YT[i]
mse_y_hat += (y_hat[i] - y)**2
mse_y_hat /= float(len(y_hat))
return mse_y_hat
import math import matplotlib.pyplot as plt from numpy import zeros, float32 from windml.datasets.nrel import NREL from windml.mapping.power_mapping import PowerMapping from sklearn.neighbors import KNeighborsRegressor from sklearn.metrics import mean_squared_error # get windpark and corresponding target. forecast is for the target turbine park_id = NREL.park_id['tehachapi'] windpark = NREL().get_windpark(park_id, 3, 2004, 2005) target = windpark.get_target() # use power mapping for pattern-label mapping. feature_window, horizon = 3, 3 mapping = PowerMapping() X = mapping.get_features_park(windpark, feature_window, horizon) y = mapping.get_labels_turbine(target, feature_window, horizon) # train roughly for the year 2004, test roughly for the year 2005. train_to, test_to = int(math.floor(len(X) * 0.5)), len(X) # train and test only every fifth pattern, for performance. train_step, test_step = 5, 5 X_train = X[:train_to:train_step] y_train = y[:train_to:train_step] X_test = X[train_to:test_to:test_step] y_test = y[train_to:test_to:test_step] # initialize and fit a KNN regressor from sklearn with k=10 neighbors. reg = KNeighborsRegressor(10, 'uniform').fit(X_train, y_train) # run the knn regression
import time
import csv
from windml.datasets.nrel import NREL
from windml.mapping.power_mapping import PowerMapping
from sklearn import preprocessing
from sklearn.preprocessing import MinMaxScaler
from sklearn.metrics import mean_squared_error
start = time.clock()
park_id = NREL.park_id['cheyenne']
windpark = NREL().get_windpark(park_id, 5, 2004,2006)
target = windpark.get_target()
feature_window, horizon = 3, 3
mapping = PowerMapping()
data_1 = np.array(mapping.get_features_park(windpark, feature_window, horizon))
data_2 = np.array(mapping.get_labels_turbine(target, feature_window, horizon)).reshape(-1,1)
#目标多向前取feature_window个点
#data1 = np.array(mapping.get_features_park(windpark, feature_window*2, horizon))
#data2 = np.array(mapping.get_labels_turbine(target, feature_window*2, horizon)).reshape(-1,1)
#region = int(data1.shape[1]/(feature_window*2))
#for i in range(region-1):
# for j in range(feature_window):
# data1=np.delete(data1,[(i+1)*feature_window],axis = 1)
#
#
lendata=len(data_1)
data1 = data_1[:lendata:3]
from windml.datasets.nrel import NREL from windml.mapping.power_mapping import PowerMapping from sklearn.grid_search import GridSearchCV from sklearn import linear_model # get windpark and corresponding target. forecast is for the target windmill park_id = NREL.park_id['tehachapi'] windpark = NREL().get_windpark(park_id, 3, 2004, 2005) target = windpark.get_target() # use power mapping for pattern-label mapping. Feature window length is 3 time # steps and time horizon (forecast) is 3 time steps. feature_window = 3 horizon = 3 mapping = PowerMapping() X = mapping.get_features_park(windpark, feature_window, horizon) Y = mapping.get_labels_mill(target, feature_window, horizon) # train roughly for the year 2004. train_to = int(math.floor(len(X) * 0.5)) # test roughly for the year 2005. test_to = len(X) # train and test only every fifth pattern, for performance. train_step, test_step = 5, 5 # fitting the pattern-label pairs reg = linear_model.LinearRegression() reg = reg.fit(X[0:train_to:train_step], Y[0:train_to:train_step])