def pointQuery(api, attribute, sort=False, reverse=False):
# result = []
result = api.liststreamkeyitems(att_dict[attribute[0]], attribute[1:],
False, MAX_RESULT)
result = getData(result["result"])
temp = []
# if attribute[0] == 'T' or attribute[0] == 'I':
# for uid in getData(api.liststreamkeyitems(
# att_dict[attribute[0]], attribute[1:], False, MAX_RESULT)["result"]):
# # print("Uid", uid)
# result += getData(api.liststreamkeyitems(UID,
# uid, False, MAX_RESULT)["result"])
# else:
# result += getData(api.liststreamkeyitems(
# att_dict[attribute[0]], attribute[1:], False, MAX_RESULT)["result"])
if attribute[0] == 'U' or attribute[0] == 'R':
for line in result:
node, _, RID = line.split(" ")[1:4]
nrid = node + 'R' + RID
RIDResult = getData(
api.liststreamkeyitems(NRID, nrid, False,
MAX_RESULT)["result"])
temp += RIDResult
# for r in RIDResult:
# if r.split(" ")[1] == node:
# temp += [r]
result += temp
if DO_VALIDATION: # and attribute[0] != 'U' and attribute[0] != 'R':
if database.validate(result, attribute, True) is False:
print("Wrong!")
return result
def problem_2_1(max_M=15):
sse_M = []
l_2_norm = []
for i in range(max_M):
x,y = getData('curvefitting.txt')
w_ols = max_likelihood(x,y,i)
bishop_plot(x,y,w_ols,'maximum likelihood, M = %s' % i)
plt.savefig('problem2_ml_M_%s.png' % i)
sse_M.append(sse(x,y,w_ols))
l_2_norm.append(sum([ii**2 for ii in w_ols])**0.5)
plt.close()
plt.plot(np.array(range(max_M)), np.array(sse_M), label='Sum of Squared Errors')
plt.savefig('problem2_OLS_SSEvsM.png')
plt.close()
plt.plot(np.array(range(max_M)), np.array(l_2_norm), label='L2 Norm of Weight Vector')
plt.savefig('problem2_OLS_L2_NormvsM.png')
plt.close()
x,y = getData('curvefitting.txt')
indices = np.random.choice(max(x.size, y.size)+1, 2)
for ii in np.sort(indices)[::-1]:
x = np.delete(x,ii)
y = np.delete(y,ii)
for M in [0,1,3,9]:
w_ols = max_likelihood(x,y,M)
bishop_plot(x,y,w_ols,'maximum likelihood w/missing, M = %s' % M)
plt.savefig('problem2_mlmissing_M_%s.png' % M)
def problem_3_2(A_or_B, plot_individual=False):
train_x, train_y = getData('regress%s_train.txt' % A_or_B)
valid_x, valid_y = getData('regress_validate.txt')
lams = np.linspace(0, 10, 101)
colors = ['b', 'g', 'r', 'c', 'm', 'y', 'k', '#808080', '#DDDD00', '#800000']
best_lams = []
best_sses = []
for M in range(10):
errors = []
for lam in lams:
error = problem_3_2_case(train_x, train_y, valid_x, valid_y,
M, lam, A_or_B, plot_individual)
errors.append(error)
if not plot_individual:
plt.plot(lams, errors, color=colors[M],
label='M=%s' % M)
best_lams.append(np.argmin(errors)/10.)
best_sses.append(min(errors))
print best_lams
print best_sses
if not plot_individual:
plt.ylim(0, 30 if A_or_B == 'A' else 100)
plt.title('Ridge regression error by $M$ and $\lambda$, training set %s' % A_or_B)
plt.xlabel('$\lambda$')
plt.ylabel('sum of squared error (against validation data)')
plt.legend()
plt.savefig('problem3_2_train%s_meta.png' % A_or_B)
plt.show()
plt.close()
def rangeQuery(api, start, end):
TIMESTAMPE = 'Timestamp'
# print(start, end)
result = []
for ts in range(start // SCALE + 1, end // SCALE):
# print('ts:', ts)
# temp = api.liststreamkeyitems(TIMESTAMPE, str(ts))['result']
temp = pointQuery(api, TIMESTAMPE, str(ts))
# print(temp)
if temp:
result += temp
# print(result)
temp = api.liststreamkeyitems(TIMESTAMPE, str(start // SCALE))['result']
# print('in range:\n', *result, sep='\n')
if temp:
data = getData(temp)
# print(data)
sl = SortedList(data, key=lambda a: a.split(" ")[0])
result += list(sl.irange(str(start), str(end)))
temp = api.liststreamkeyitems(TIMESTAMPE, str(end // SCALE))['result']
if temp:
data = getData(temp)
sl = SortedList(data, key=lambda a: a.split(" ")[0])
# print('end:')
# print(*list(sl.irange(maximum=str(end))), sep='\n')
result += list(sl.irange(str(start), str(end)))
return result
def pointQuery(api, attribute, sort=False, reverse=False):
result = api.liststreamkeyitems(att_dict[attribute[0]], attribute[1:],
False, MAX_RESULT)
if DO_VALIDATION:
if database.validate(getData(result["result"]), attribute,
True) == False:
print("Wrong!")
return getData(result["result"])
def pointQuery(api, attribute, sort=True, reverse=False):
result = api.liststreamkeyitems(
att_dict[attribute[0]], attribute[1:], False, FILE_SIZE)
if DO_VALIDATION:
validate(getData(result["result"]), attribute[1:])
result = getData(result["result"])
att_name = att_dict[attribute[0]]
if type(ATTRIBUTE_TYPE[att_name]) is int and sort:
result = sortResult(result, att_name, reverse)
# print(*result, sep='\n')
# input()
return result
def pointQuery(api, attribute, value):
result = getData(
api.liststreamkeyitems(attribute, value, False, MAX_RESULT)["result"])
temp = []
if attribute == "User" or attribute == "Resource":
for line in result:
node, ID = line.split(" ")[1:3]
RIDResult = getData(
api.liststreamkeyitems('Ref-ID', ID, False,
MAX_RESULT)["result"])
for r in RIDResult:
if r.split(" ")[1] == node:
temp += [r]
result += temp
return result
def pointQuery(api, stream, key):
if stream == "Timestamp":
txids = getTXids(api, PREFIX + str(SCALE), str(int(key) // SCALE))
# resulls = api.getstreamitem(DATA, txids[0])
elif stream == DATA or stream == 'Activity':
return getData(
api.liststreamkeyitems(stream, key, False, MAX_RESULT)["result"])
else:
txids = getTXids(api, stream, key)
args = [[DATA, txid] for txid in txids]
results = api.batch('getstreamitem', args)
result = []
for r in results:
# print(r["result"])
data = bytes.fromhex(r["result"]["data"]).decode(ENCODE_FORMAT)
if stream == "Timestamp":
if data.split(" ")[0] == key:
result = [data]
break
# return data
else:
result.append(data)
# print(result)
# input()
# if stream == 'ts1':
# stream = "Timestamp"
# if database.validate(result, stream, key, True) is False:
# print("Wrong!")
return result
def update_metrics(n):
data = util.getData()
twitter_before = data['original']
twitter_after = data['text']
session = data['session']
#print(twitter_before)
timenow = [util.convertime(i[0]) for i in session]
his = [i[1] for i in session]
sessionnow = dict(zip(timenow, his))
# print(timenow)
# print(his)
style = {'padding': '5px', 'fontSize': '25px'}
result = []
result.append(
html.Span('Before PreProcess: ' + twitter_before, style=style))
result.append(html.Div())
result.append(html.Span('After PreProcess: ' + twitter_after, style=style))
result.append(html.Div())
for key, value in sessionnow.items():
result.append(
html.Span('History Session : ' + str(key) + str(value),
style=style))
result.append(html.Div())
return result
def main():
X, Y = getData()
X, Y = shuffle(X, Y)
K = len(np.unique(Y))
N = len(Y)
T = np.zeros((N, K))
for i in range(N):
T[i, Y[i]] = 1 # one hot encoding for targets
batch_sz = 500
learning_rate = [10e-5, 10e-6, 10e-7, 10 - 8, 10e-9]
num_batches = len(learning_rate)
trainCost = []
validCost = []
accValid = []
accTrain = []
for i in range(num_batches):
m = nn.NeuralNetwork(numHiddenLayer=1,
numHiddenUnits=200,
actFunc="Tanh")
trainCost[i], validCost[i], accTrain[i], accValid[i] = m.train(
X, T, epochs=10000, learning_rate=10e-7, reg=10e-7)
print("Final Train Accuracy {}".format(accTrain))
print("Final Valid Accuracy {}".format(accValid))
legend1, = plt.plot(trainCost, label='training error')
legend2, = plt.plot(validCost, label='validation error')
plt.legend([legend1, legend2])
plt.show()
def calEmotion(filename, city):
# load dict
most, very, more, insuff, ish, posdict, negdict, over = dict.init_dict()
jieba.load_userdict("dic/stock_dict.txt")
# init score
score = {}
cur_date = 0
review_id = 0
#init excel hander
w, ws = createExcel()
# open database & get data
rawdata, conn, cur = util.getData(filename, city)
for row in rawdata:
# collect tokens from yield
seg_list, cur_date, review = getTokens(score, row, cur_date)
#use sentiment score to cal the emotion by one comment and add it to score.
result = dict.sentiment_score_list(seg_list, posdict, negdict, most,
very, more, ish, insuff, over)
score[cur_date].append(result)
# write down the review to xls
review_id = review_id + 1
ws.write(review_id, 0, review)
ws.write(review_id, 1, result)
w.save("excel/" + filename + ".xls")
# close the connection with database
util.closeDB(conn, cur)
# generate the emotionlist from score
emotionList = genEmotionList(score)
# show pic of the stock price and emotion by one stock
showPriceAndEmotion(emotionList, filename, city)
def main():
X, Y = getData(balance_ones=False)
# 48by48 = 2304 dimensions per sample
# Data structure
# 0-48 one row
# 48-96 second row and so on
# infinite loop. Break out of it by via input
while True:
# loop through 7 emotions
for i in range(7):
x, y = X[Y == i], Y[
Y ==
i] # choose all the data points that equal to this emotion
N = len(y) # number of data points equals to this emotion
j = np.random.choice(N) # randomly select this point
plt.imshow(x[j].reshape(48, 48),
cmap='gray') # plot this point. Reshape it
plt.title(label_map[y[j]]) # plot the label - emotion
plt.show()
prompt = input('Quit? Enter Y:\n')
if prompt == 'Y':
break
def main():
Xtrain, Ytrain, Xvalid, Yvalid = getData()
model = ANN(200)
model.fit(Xtrain, Ytrain, Xvalid, Yvalid, reg=0, show_fig=True)
print(model.score(Xvalid, Yvalid))
def pointQuery(api, attribute, display=False, validation=DO_VALIDATION):
result = api.liststreamkeyitems(STREAM, attribute, False, FILE_SIZE)
if validation:
validate(getData(result["result"]), attribute[1:])
if display:
display(result["result"])
return result["result"]
def calEmotion(filename,city):
# load dict
most,very,more,insuff,ish,posdict,negdict,over=dict.init_dict()
jieba.load_userdict("dic/stock_dict.txt")
# init score
score = {}
cur_date = 0
review_id =0
#init excel hander
w, ws = createExcel()
# open database & get data
rawdata, conn, cur = util.getData(filename,city)
for row in rawdata:
# collect tokens from yield
seg_list,cur_date, review = getTokens(score,row,cur_date)
#use sentiment score to cal the emotion by one comment and add it to score.
result = dict.sentiment_score_list(seg_list,posdict,negdict,most,very,more,ish,insuff,over)
score[cur_date].append(result)
# write down the review to xls
review_id = review_id +1
ws.write(review_id,0,review)
ws.write(review_id,1,result)
w.save("excel/" + filename + ".xls")
# close the connection with database
util.closeDB(conn, cur)
# generate the emotionlist from score
emotionList = genEmotionList(score)
# show pic of the stock price and emotion by one stock
showPriceAndEmotion(emotionList,filename,city)
def problem_3_1():
x, y = getData('curvefitting.txt')
for M in [1, 2, 3]:
for lam in [0, 0.0003, 0.001, 0.003]:
w_ridge = polynomial_ridge(x, y, M, lam)
bishop_plot(x, y, w_ridge, 'ridge regression, M=%s, lam=%s' % (M, lam))
plt.savefig('problem3_1_M_%s_lam_%s.png' % (M, lam))
print 'just tried M = %s, lam=%s' % (M, lam)
print ' sse was %s' % sse(x, y, w_ridge)
def rangeQuery(api, start, end):
result = []
stream = att_dict['T']
timestamps = api.liststreamkeys(stream)["result"]
sl = SortedList(list(map(int, [key['key'] for key in timestamps])))
for timestamp in sl.irange(start, end):
result += getData(
api.liststreamkeyitems(stream, str(timestamp))['result'])
return result
def rangeQuery(api, start, end):
TIMESTAMPE = 'Timestamp'
result = []
for ts in range(start // SCALE + 1, end // SCALE):
temp = pointQuery(api, TIMESTAMPE, str(ts))
if temp:
result += temp
temp = api.liststreamkeyitems(TIMESTAMPE, str(start // SCALE))['result']
if temp:
data = getData(temp)
sl = SortedList(data, key=lambda a: a.split(" ")[0])
result += list(sl.irange(str(start), str(end)))
temp = api.liststreamkeyitems(TIMESTAMPE, str(end // SCALE))['result']
if temp:
data = getData(temp)
sl = SortedList(data, key=lambda a: a.split(" ")[0])
result += list(sl.irange(str(start), str(end)))
return result
def andQuery(api, attributes, display=False):
resultSet = []
for attr in attributes:
# print(getData(pointQuery(api, attr)))
resultSet.append(set(getData(pointQuery(api, attr))))
result = resultSet[0]
for i in range(1, len(resultSet)):
result &= resultSet[i]
if display:
display(result)
def __init__(self, inputDir, modelDir,action, **kwargs):
self.action = action
self.modelDir=modelDir
self.inputDir=inputDir
self.clf =None
if(self.action=='training'):
self.mats1, self.spams1, self.mats2, self.spams2 =util.getData(self.inputDir)
else:
self.mats2, self.spams2 =util.getPredictData(self.inputDir)
self.mats1, self.spams1 = None , None
self.clf = joblib.load(self.modelDir)
def update_graph(n):
data = util.getData()
# most recent session
session = data['session'][-1]
time = session[0]
count = session[1]
num = len(count.keys())
sentiment_number = []
for i in count.keys():
sentiment_number.append((count[i][0], count[i][1]))
labels = ['Positive', 'Negative']
# generate singel pie format
piedata = []
for i in range(num):
onepie = util.PieOneData(
[sentiment_number[i][0], sentiment_number[i][1]], labels, i)
piedata.append(onepie)
# generate annotations
annotations = []
# position:
if num == 1:
position = [0.5]
elif num == 2:
position = [0.2, 0.8]
elif num == 3:
position = [0.15, 0.5, 0.85]
elif num == 4:
position = [0.11, 0.375, 0.625, 0.89]
tmp = 0
names = []
for i in count.keys():
annotations.append(util.subtitle(i, position[tmp]))
names.append(i)
tmp += 1
title = " -- ".join(names)
#print(piedata)
#print(annotations)
layout = util.layout(annotations, num, title)
fig = util.PieAllData(piedata, layout)
return fig
def pointQuery(api, attribute, value):
if attribute == "Timestamp":
TSResult = getData(api.liststreamkeyitems(
'Timestamp', str(int(value) // SCALE), False, MAX_RESULT)["result"])
for tsr in TSResult:
ts = tsr.split(" ")[0]
if ts == value:
return [tsr]
result = getData(api.liststreamkeyitems(
attribute, value, False, MAX_RESULT)["result"])
temp = []
if attribute == "User" or attribute == "Resource":
for line in result:
node, ID = line.split(" ")[1:3]
RIDResult = getData(api.liststreamkeyitems(
'Ref-ID', ID, False, MAX_RESULT)["result"])
for r in RIDResult:
if r.split(" ")[1] == node:
temp += [r]
result += temp
return result
def main():
X, Y = getData(balance_ones=False)
while True:
for i in xrange(len(label_map)):
x, y = X[Y==i], Y[Y==i]
N = len(y)
j = np.random.choice(N)
plt.imshow(x[j].reshape(48,48), cmap='gray')
plt.title(label_map[y[j]])
prompt = raw_input('Quit? Enter Y:\n')
if prompt == 'Y':
break
def main():
X, Y = getData(balance_ones=False)
while (True):
for i in range(7):
x, y = X[Y == i], Y[Y == i]
N = len(y)
j = np.random.choice(N)
plt.imshow(x[j].reshape(48, 48), cmap='gray')
plt.title(labels[y[j]])
plt.show()
prompt = input("Quit the program? Y/N\n")
if prompt == 'Y':
break
def main():
X, Y, _, _ = getData(balance_ones=False)
while True:
for i in range(7):
x, y = X[Y == i], Y[Y == i]
N = len(y)
j = np.random.choice(N)
plt.imshow(x[j].reshape(48, 48), cmap='gray')
plt.title(label_map[y[j]])
plt.show()
prompt = input('Quit? Enter Y:\n')
if prompt.lower().startswith('y'):
break
def main():
X, Y = getData(balance_ones=False)
while True:
for i in xrange(7):
x, y = X[Y==i], Y[Y==i]
N = len(y)
j = np.random.choice(N)
plt.imshow(x[j].reshape(48, 48), cmap='gray')
plt.title(label_map[y[j]])
plt.show()
prompt = raw_input('Quit? Enter Y:\n')
if prompt == 'Y':
break
def pointQuery(api, stream, key):
if stream == "Timestamp":
txids = getTXids(api, PREFIX + str(SCALE), str(int(key) // SCALE))
# resulls = api.getstreamitem(DATA, txids[0])
elif stream == DATA or stream == 'Activity':
return getData(
api.liststreamkeyitems(stream, key, False, MAX_RESULT)["result"])
else:
txids = getTXids(api, stream, key)
args = [[DATA, txid] for txid in txids]
results = api.batch('getstreamitem', args)
result = []
for r in results:
# print(r["result"])
data = bytes.fromhex(r["result"]["data"]).decode(ENCODE_FORMAT)
if stream == "Timestamp":
if data.split(" ")[0] == key:
result = [data]
break
# return data
else:
result.append(data)
if stream == "User" or stream == 'Resource':
# txids = getTXids(api, stream, key)
nrids = []
for line in result:
node, _, rid = line.split(" ")[1:4]
nrids.append(node + 'R' + rid)
args = [[NRID, nrid, False, MAX_RESULT] for nrid in nrids]
temp = api.batch('liststreamkeyitems', args)
# print(*temp, sep='\n')
# input()
ridResults = []
for r in temp:
ridResults += r["result"]
txids = [tx["txid"] for tx in ridResults]
args = [[DATA, txid] for txid in txids]
results = api.batch('getstreamitem', args)
temp = []
for r in results:
data = bytes.fromhex(r["result"]["data"]).decode(ENCODE_FORMAT)
temp.append(data)
result += temp
# if database.validate(result, stream, key, True) is False:
# print("Wrong!")
return result
def run_standalone(c, w, o):
config, weights = loadModelFromDisk(c, w)
x_train, y_train = getData()
if x_train is None or y_train is None:
print("Data could not be loaded, abort.", file=sys.stderr)
sys.exit(1)
else:
print("Data loaded")
print("Starting training")
new_weights = trainModel(config, weights, x_train, y_train)
print("Finished training")
writeUpdatesToDisk(new_weights, o)
def main():
t0 = datetime.now()
print("now: ", t0)
X, Y = getData()
dt = datetime.now() - t0
print("dt: ", dt)
M = 10 # hidden units
model = ANN(M)
model.fit(X, Y, show_fig=True)
print(model.score(X, Y))
def problem_2(M):
x,y = getData('curvefitting.txt')
w_ols = max_likelihood(x, y, M)
print w_ols
low_limit_x = min(x) - 0.1*abs(max(x)-min(x))
hi_limit_x = max(x) + 0.1*abs(max(x)-min(x))
x_points = np.linspace(low_limit_x, hi_limit_x, 100)
y_hat = np.array([compute_yhat(pt, w_ols) for pt in x_points])
low_limit_y = min(y) - 0.2*abs(max(y) - min(y))
hi_limit_y = max(y) + 0.2*abs(max(y)-min(y))
plt.plot(x_points, y_hat, '-b', label='Max Likelihood, M='+str(M))
plt.scatter(x, y, label='Data')
plt.plot(x_points, np.sin(2*np.pi*x_points), '-', color='0.5', label='Sin(2*pi*x)')
plt.xlim(low_limit_x, hi_limit_x)
plt.ylim(low_limit_y, hi_limit_y)
plt.legend()
plt.show()
plt.savefig('Problem2_M_'+str(M)+'.png')
def pointQuery(api, attribute, sort=False, reverse=False):
pointers = getPointers(api, attribute)
result = []
# print(pointers)
# input()
# api.batch(api.liststreamkeyitems, [
# (DATA, p, False, MAX_RESULT) for p in pointers])
# for p in pointers:
# result += getValue(api, p)
args = [[DATA, p, False, MAX_RESULT] for p in pointers]
results = api.batch('liststreamkeyitems', args)
for r in results:
result.append(getData(r["result"])[0])
# print(result)
# if database.validate(result, attribute, True) is False:
# print("Wrong!")
# input()
return result
def main():
X, Y, Xtest, Ytest = getData()
print(Y)
while True:
for i in range(7): # loop through labels
x, y = X[Y == i], Y[
Y == i] # select data that corresponds to this label
print("x, y: ", x, y)
N = len(y)
print("N: ", N)
j = np.random.choice(
N) # select a random data point (index) (image)
plt.imshow(x[j].reshape(48, 48), cmap='gray')
plt.title(label_map[y[j]])
plt.show()
prompt = input('Y to quit:')
if prompt == 'Y':
break
def main():
X, Y = getData()
model = LogisticModel()
model.fit(X, Y)
print(model.score(X, Y))
def main():
X, Y = getData()
# X, Y = getBinaryData()
model = ANN([2000, 1000])
model.fit(X, Y, show_fig=True)
# -*- coding:utf-8 -*- import codecs import numpy as np from util.getData import * from regression.standRegression import * ''' train_data:xtrain train_label:ytrain test_data:xpred test_time:date_time ''' getdata = getData() # 0 1 2 3 4 5 6 7 8 # :return: data [星期, 天气, 最高温, 最低温, 节假日长度, 第i个节假日, 工作日长度, 第i个工作日, date, # 9 10 11 # 小时, 线路, 数量(y)] # :return: 返回20150101到20150107的测试数据集 # 0 1 2 3 4 5 6 7 8 9 # [日期, 星期, 天气, 最高温, 最低温, 节假日长度, 第i个节假日, 工作日长度, 第i个工作日, 小时] # # """ features = [1, 2, 3, 4, 5, 7, 9] data = getdata.get_train_data(min_day="20140901", max_day="20141231", line_num=[6]) train = np.array(data)
def main():
X, Y = getData()
model = ANN(200)
model.fit(X, Y, reg=0, show_fig=True)
print(model.score(X, Y))
import requests
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.dates as dts
from keras.models import Sequential
from keras.layers import Dense, Dropout
from util import getData, processData
np.random.seed(7)
if __name__ == "__main__":
data = getData('2012-01-10', '2017-08-25')
X_train, X_test, Y_train, Y_test = processData(data)
print(X_train)
# nn = NN([(39, 500)], [500, 1])
# nn.train(X_train, Y_train, X_test, Y_test, 10000)
model = Sequential()
model.add(Dense(500, input_shape=(19, ), init='uniform',
activation='relu'))
# model.add(Dense(500, init='uniform', activation='relu'))
model.add(Dropout(0.25))
model.add(Dense(400, init='uniform', activation='relu'))
model.add(Dropout(0.25))
model.add(Dense(250, init='uniform', activation='relu'))
model.add(Dropout(0.25))
model.add(Dense(1, init='uniform', activation='linear'))
import numpy as np import sys import json import util from sklearn.linear_model import LogisticRegression C=1.1 mats1, spams1, mats2, spams2 =util.getData(sys.argv[1]) for i in range(3): print C clf = LogisticRegression(C=C).fit(mats1, spams1) predict1 = clf.predict(mats1) predict2 = clf.predict(mats2) print util.validation(predict1, spams1), util.validation(predict2, spams2) C=C+0.5
def main():
X, Y = getData()
model = LogisticModel()
model.fit(X, Y, show_fig=True)
print(model.score(X, Y))
def main():
Xtrain, Ytrain, Xvalid, Yvalid = getData()
print("read data..........")
model = ANN([2000, 1000, 500])
model.fit(Xtrain, Ytrain, Xvalid, Yvalid, show_fig=True)
# -*- coding:utf-8 -*- __author__ = 'HeYongxing' import numpy as np from matplotlib.pylab import show import matplotlib.pyplot as plt from mpl_toolkits.mplot3d import Axes3D from util.getData import * # 十号线路 # [星期, 天气, 最高温, 最低温, 小时, 线路, 数量(y), datetime] getData = getData() data = getData.get_train_data(min_day="20140901", max_day="20141231", line_num=[15]) train = np.array(data) xtrain = train[:, 0:-3] xdate = train[:, -1] ytrain = train[:, -2] # fig = figure() # ax1 = fig.add_subplot(1, 1, 1) # ax1.scatter(xtrain[:, 0], ytrain, c='red', # alpha=0.3, edgecolors='none') # ax1.set_xlabel(u'星期') # ax1.grid(True) # show() # # # fig = figure()
from random import shuffle
from flask import (Flask, request, make_response, jsonify, redirect)
from util import getData, getData2
import os
app = Flask(__name__)
songs, movies = getData()
_t = getData2()
songs.update(_t)
song_indices = {}
for v in songs.itervalues():
for each_song in v:
index = str(each_song[1].split()[-1][1:-1])
label = ' '.join(each_song[1].split()[:-1])
song_indices[index] = label
movie_indices = {}
for v in movies.itervalues():
for each_movie in v:
index = str(each_movie[1].split()[-1][1:-1])
label = ' '.join(each_movie[1].split()[:-1])
movie_indices[index] = label
@app.route('/')
def index():
return make_response(open('templates/index.html').read())
def main():
X, Y = getData()
model = ANN(200)
model.fit(X, Y, reg=0, show_fig=True)
print(model.score(X, Y))
def main():
X, Y = getData()
# X, Y = getBinaryData()
model = ANN([2000, 1000, 500])
model.fit(X, Y, show_fig=True)
Calculates walking distances between a random sampling of buildings and compares to optimal
straight-line path
Sam Shen
"""
import json, util, sys
import networkx as nx
from decimal import *
from osmread import Node
FIRST_RUN = False
SAMPLE_REPEAT_SIZE = 3 #the number of times to sample
SAMPLE_SIZE_SQRT = 8 #the squareroot of the sample size
graph, nodes = util.init_graph(util.getData('../data/berkeley_map.osm'))
print("Initialized graph...")
buildings = json.load(open("../data/buildings.json"))
#only have to run this part once
if FIRST_RUN:
util.batch_discretize(nodes, buildings, '../data/discrete_locs.txt')
str_coords = [] #list of string coordinate pairs
dec_coords = set() #set of decimal coordinate pairs
discrete_nodes = [] #the nodes that represent discrete buildings
try:
file = open("../data/discrete_locs.txt")
except:
print("Did you download discrete_locs.txt OR run batch discretizing by setting FIRST_RUN to true?")