def predict(netWork, DS, preData):
trainer = BackpropTrainer(netWork, DS, verbose=True, learningrate=0.01)
#maxEpochs 即需要的最大收敛迭代次数
trainer.trainUntilConvergence(maxEpochs=1000)
#获取预测结果
preValue = n.activate(preData)
#设置预测值
if preValue > 0:
v = 1
else:
v = -1
#将预测结果保存到数据库中
g.save(v, time)
def predict(netWork,DS,preData): trainer = BackpropTrainer(netWork,DS,verbose = True,learningrate = 0.01) #maxEpochs 即需要的最大收敛迭代次数 trainer.trainUntilConvergence(maxEpochs = 1000) #获取预测结果 preValue = n.activate(preData) #设置预测值 if preValue > 0: v = 1 else: v = -1 #将预测结果保存到数据库中 g.save(v,time)
result = {}
client = client_factory('CAISO')
for day in dayList:
temp = getData.get_daily_data(year, month, day, client)
if firstIteration:
firstIteration = False
for category in temp:
result[category] = []
# Add one more day in the final result
result = getData.concatenate_day(result, temp)
time.sleep(1)
# Save to csv
path = os.path.join(os.path.dirname(__file__), 'csv_result', str(year) + '-' + str(month) + '.csv')
getData.write_data_as_csv(result, path)
# Plot data
plt.figure(figsize=(20, 20), dpi=80)
nbPlot = 100 * (len(result) - 1) + 10
for category in result:
if not category == 'time':
nbPlot += 1
plt.subplot(nbPlot)
plt.plot(result['time'], result[category], label=category)
plt.xlabel('Time')
plt.ylabel('Power (MW)')
plt.legend()
path = os.path.join(os.path.dirname(__file__), 'graph_result', str(year) + '-' + str(month))
getData.save(path, ext='png', close=True, verbose=True)