def model2(row):
"""
MODEL2 : binary day and week attributes + holiday (binary)
indeksi : 0-6 dnevi, 7-31 ura
server: 152.32996
lokalno: 129.303771
"""
result = np.zeros(7 + 24 + 2)
day = lpputils.parsedate(row[DEP_IDX]).weekday()
hour = lpputils.parsedate(row[DEP_IDX]).hour
result[day] = 1
result[7 + hour] = 1
date = lpputils.parsedate(row[DEP_IDX]).date()
holiday = 0
school_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
result[-2] = holiday
result[-1] = school_hol
return result
def zgradi_matrike(linija, training):
d = open("prazniki_in_dela_prosti_dnevi.csv", "rt", encoding="latin1")
branje = csv.reader(d)
next(branje)
prazniki = []
for d in branje:
prazniki.append(
datetime.datetime.strptime((d[0].split(";", 1))[0],
"%d.%m.%Y").date())
if training:
x = []
y = []
for d in linija:
x.append(napolni_x(d, prazniki))
y.append(
lpputils.tsdiff(lpputils.parsedate(d[-3]),
lpputils.parsedate(d[-1])))
X = linear.append_ones(np.array(x))
Y = np.array(y)
return X, Y
else:
x = []
originalen_datum = []
dejanski_cas = []
route = []
for d in linija:
originalen_datum.append(d[-3])
dejanski_cas.append(d[-1])
route.append(d[3])
x.append(napolni_x(d, prazniki))
X = linear.append_ones(np.array(x))
return route, dejanski_cas, originalen_datum, X
def model4(row):
"""
MODEL4 : binary day and week attributes + all holiday (binary)
indeksi : 0-6 dnevi, 7-31 ura, pocitnice 3x
server: 150.86627
lokalno: 128.68312
"""
result = np.zeros(7 + 24 + 3)
day = lpputils.parsedate(row[DEP_IDX]).weekday()
hour = lpputils.parsedate(row[DEP_IDX]).hour
result[day] = 1
result[7 + hour] = 1
date = lpputils.parsedate(row[DEP_IDX]).date()
holiday = 0
school_hol = 0
summer_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
if lpputils.parsedate(SUMMER_HOL[0]).date() <= date <= lpputils.parsedate(
SUMMER_HOL[1]).date():
summer_hol = 1
result[-3] = summer_hol
result[-2] = holiday
result[-1] = school_hol
return result
def model1(row):
"""
MODEL1 : norm day, norm hour, holiday, school holiday, avg. driver, avg. bus
"""
result = np.zeros(6)
result[0] = lpputils.parsedate(row[DEP_IDX]).weekday() / 7.0 # day
result[1] = lpputils.parsedate(row[DEP_IDX]).hour / 24.0 # hour
date = lpputils.parsedate(row[DEP_IDX]).date()
holiday = 0
school_hol = 0
#summer_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
if lpputils.parsedate(SUMMER_HOL[0]).date() <= date <= lpputils.parsedate(
SUMMER_HOL[1]).date():
summer_hol = 1
result[2] = holiday
result[3] = school_hol
#result[4] = summer_hol
result[4] = driver_average(row[DRV_IDX])
result[5] = bus_average(row[BUS_IDX])
return result
def visualize(train_data, _month, day_s, day_e):
"""
Function which outputs daily travel time by hour (graph + text)
"""
comp_data = []
times = np.zeros(24)
cnts = np.zeros(24)
for d in range(day_s, day_e, 1):
for row in train_data:
date = lpputils.parsedate(row[DEP_IDX])
hour = date.hour
month = date.month
day = date.day
if month == _month and day == d:
times[hour] += lpputils.tsdiff(row[ARR_IDX], row[DEP_IDX])
cnts[hour] += 1
norm_times = [
float(times[i]) / (float(cnts[i]) + 0.0000000000001)
for i in range(len(times))
]
#comp_data.append(np.asarray(norm_times))
print(norm_times)
#with open('vizualizacija.csv', 'wb') as abc:
# np.savetxt(abc, np.asarray(comp_data), delimiter=",", fmt="%d")
#data = np.genfromtxt('vizualizacija.csv', delimiter=',')
#for i in range(len(data)):
plt.plot(norm_times, label='the data')
plt.show()
def model3(row):
"""
MODEL3 : binary day and week attributes + drivers + buses + holiday
indeksi : dnevi, ure, vozniki, busi, pocitnice
server: ???
lokalno: 129.2525
"""
day_offset = 0
hour_offset = 7
driver_offset = 31
buses_offset = driver_offset + len(driver_idxs)
holiday_offset = buses_offset + len(buses_idxs)
result = np.zeros(holiday_offset + 3)
day = lpputils.parsedate(row[DEP_IDX]).weekday()
hour = lpputils.parsedate(row[DEP_IDX]).hour
result[day_offset + day] = 1
result[hour_offset + hour] = 1
if row[DRV_IDX] in driver_idxs.keys():
result[driver_offset + driver_idxs[row[DRV_IDX]]] = 1
if row[BUS_IDX] in buses_idxs.keys():
result[buses_offset + buses_idxs[row[BUS_IDX]]] = 1
date = lpputils.parsedate(row[DEP_IDX]).date()
holiday = 0
school_hol = 0
summer_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
if lpputils.parsedate(SUMMER_HOL[0]).date() <= date <= lpputils.parsedate(
SUMMER_HOL[1]).date():
summer_hol = 1
result[-3] = summer_hol
result[-2] = holiday
result[-1] = school_hol
return result
def model_to_csv(train_data, filename):
file = open(filename, 'w')
file.write('line,month,day,hour,travel\n')
for row in train_data:
date = lpputils.parsedate(row[DEP_IDX])
hour = date.hour
day = date.weekday()
month = date.month
travel = lpputils.tsdiff(row[ARR_IDX], row[DEP_IDX])
file.write('{0},{1},{2},{3},{4}\n'.format(row[2], month, day, hour,
travel))
def model5(row):
"""
MODEL5 : binary day and week attributes + all holiday (binary)
indeksi : 0-6 dnevi, 7-31 ura, pocitnice 3x, padavine
server: 184.51330
lokalno: 147.68
"""
global arso
result = np.zeros(7 + 24 + 4)
day = lpputils.parsedate(row[DEP_IDX]).weekday()
hour = lpputils.parsedate(row[DEP_IDX]).hour
result[day] = 1
result[7 + hour] = 1
date = lpputils.parsedate(row[DEP_IDX]).date()
holiday = 0
school_hol = 0
summer_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
if lpputils.parsedate(SUMMER_HOL[0]).date() <= date <= lpputils.parsedate(
SUMMER_HOL[1]).date():
summer_hol = 1
result[-4] = summer_hol
result[-3] = holiday
result[-2] = school_hol
if date.strftime("%Y-%m-%d") in arso.keys():
result[-1] = arso[date.strftime("%Y-%m-%d")][0]
else:
print("{0} - No weather data !!!".format(date.strftime("%Y-%m-%d")))
return result
def model9(row):
"""
MODEL9 : binary day and hour attributes + all holiday (binary) -> added 20 min interval between 06 and 09
indeksi : 30 * 7 kombinacije dan ura, pocitnice 3x, padavine, detour
server: 179.68471
lokalno: 142.....
uporaba: rezultati 8, 9(+45 sekund na linijo 1), 10(+65 sekund na linijo 1)
"""
global arso
result = np.zeros(7 * 30 + 5)
date = lpputils.parsedate(row[DEP_IDX]).date()
day = lpputils.parsedate(row[DEP_IDX]).weekday()
hour = lpputils.parsedate(row[DEP_IDX]).hour
minutes = lpputils.parsedate(row[DEP_IDX]).minute
day_offset = 30 * day
if hour < 6:
result[day_offset + hour] = 1
elif 6 <= hour <= 8:
offset = (hour - 6) * 2
if 0 <= minutes <= 20:
result[day_offset + hour + offset] = 1
elif 20 < minutes <= 40:
result[day_offset + hour + offset + 1] = 1
elif 40 < minutes <= 59:
result[day_offset + hour + offset + 2] = 1
else:
result[day_offset + hour + 6] = 1
holiday = 0
school_hol = 0
summer_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
if lpputils.parsedate(SUMMER_HOL[0]).date() <= date <= lpputils.parsedate(
SUMMER_HOL[1]).date():
summer_hol = 1
result[-5] = summer_hol
result[-4] = holiday
result[-3] = school_hol
if date.strftime("%Y-%m-%d") in arso.keys():
result[-2] = 1 if arso[date.strftime("%Y-%m-%d")][0] > 20 else 0
else:
print("No data for weather !!!")
line = row[2]
if row[3][0:2] in ['B ', 'G ', 'I ', 'Z ']:
line += row[3][0]
result[-1] = check_detour(line, row[DEP_IDX], detours)
return result
def model8(row):
"""
MODEL8 : binary day and week attributes + all holiday (binary) + weather + detour-> added 20 min interval between 06 and 09
indeksi : 0-6 dnevi, 7-37 ura, pocitnice 3x, padavine, obvoz
server: ?
lokalno:
"""
global arso
global detours
result = np.zeros(7 + 30 + 5)
date = lpputils.parsedate(row[DEP_IDX]).date()
day = lpputils.parsedate(row[DEP_IDX]).weekday()
hour = lpputils.parsedate(row[DEP_IDX]).hour
minutes = lpputils.parsedate(row[DEP_IDX]).minute
result[day] = 1
if hour < 6:
result[7 + hour] = 1
elif 6 <= hour <= 8:
offset = (hour - 6) * 2
if 0 <= minutes <= 20:
result[7 + hour + offset] = 1
elif 20 < minutes <= 40:
result[7 + hour + offset + 1] = 1
elif 40 < minutes <= 59:
result[7 + hour + offset + 2] = 1
else:
result[7 + hour + 6] = 1
holiday = 0
school_hol = 0
summer_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
if lpputils.parsedate(SUMMER_HOL[0]).date() <= date <= lpputils.parsedate(
SUMMER_HOL[1]).date():
summer_hol = 1
result[-5] = summer_hol
result[-4] = holiday
result[-3] = school_hol
if date.strftime("%Y-%m-%d") in arso.keys():
result[-2] = arso[date.strftime("%Y-%m-%d")][0]
else:
print("No data !!!")
line = row[2]
if row[3][0:2] in ['B ', 'G ', 'I ', 'Z ']:
line += row[3][0]
result[-1] = check_detour(line, row[DEP_IDX], detours)
return result
def model6(row):
"""
MODEL6 : binary day and week attributes + all holiday (binary) + rush
indeksi : 0-6 dnevi, 7-31 ura, pocitnice 3x, rush1, rush2
server: ??
lokalno: 147.68
"""
result = np.zeros(7 + 24 + 5)
day = lpputils.parsedate(row[DEP_IDX]).weekday()
hour = lpputils.parsedate(row[DEP_IDX]).hour
result[day] = 1
result[7 + hour] = 1
date = lpputils.parsedate(row[DEP_IDX]).date()
holiday = 0
school_hol = 0
summer_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
if lpputils.parsedate(SUMMER_HOL[0]).date() <= date <= lpputils.parsedate(
SUMMER_HOL[1]).date():
summer_hol = 1
result[-5] = summer_hol
result[-4] = holiday
result[-3] = school_hol
if hour >= 3 and hour <= 6:
result[-2] = (hour % 3) / 3
if hour >= 15 and hour <= 18:
result[-1] = (3 - hour % 15) / 3
return result
def model10(row):
"""
MODEL10 : binary day and hour attributes + all holiday (binary) + drivers-> added 20 min interval between 06 and 09
indeksi : 30 * 7 kombinacije dan ura, vozniki, pocitnice 3x, padavine
server: 145.82494
lokalno: 122.3
uporaba: rezultati 11,12,13
"""
global arso
result = np.zeros(7 * 30 + len(driver_idxs) + 4)
date = lpputils.parsedate(row[DEP_IDX]).date()
day = lpputils.parsedate(row[DEP_IDX]).weekday()
hour = lpputils.parsedate(row[DEP_IDX]).hour
minutes = lpputils.parsedate(row[DEP_IDX]).minute
day_offset = 30 * day
if hour < 6:
result[day_offset + hour] = 1
elif 6 <= hour <= 8:
offset = (hour - 6) * 2
if 0 <= minutes <= 20:
result[day_offset + hour + offset] = 1
elif 20 < minutes <= 40:
result[day_offset + hour + offset + 1] = 1
elif 40 < minutes <= 59:
result[day_offset + hour + offset + 2] = 1
else:
result[day_offset + hour + 6] = 1
driver_offset = 7 * 30
if row[DRV_IDX] in driver_idxs.keys():
result[driver_offset + driver_idxs[row[DRV_IDX]]] = 1
holiday = 0
school_hol = 0
summer_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
if lpputils.parsedate(SUMMER_HOL[0]).date() <= date <= lpputils.parsedate(
SUMMER_HOL[1]).date():
summer_hol = 1
result[-4] = summer_hol
result[-3] = holiday
result[-2] = school_hol
#if date.strftime("%Y-%m-%d") in arso.keys():
# result[-1] = max([arso[date.strftime("%Y-%m-%d")][0],0])
#else:
# print("No data !!!")
return result
def napolni_x(d, prazniki):
primer = []
datum = lpputils.parsedate(d[-3])
primer.append(1 if datum.minute <= 30 else 0)
primer.append(1 if datum.minute > 30 else 0)
primer.append(1 if datum.hour < 20 and datum.hour > 6 else 0)
for ura in range(24):
primer.append(1 if datum.hour == ura else 0)
for dan in range(7):
primer.append(1 if datum.weekday() == dan else 0)
primer.append(0)
for praznik in prazniki:
if datum.date() == praznik:
primer[len(primer) - 1] = 1
break
return primer
def model7(row):
"""
MODEL7 : binary day and week attributes + all holiday (binary) -> added 20 min interval between 06 and 09
indeksi : 0-6 dnevi, 7-37 ura, pocitnice 3x, padavine
server: ?
lokalno: 128.68
"""
global arso
result = np.zeros(7 + 30 + 4)
date = lpputils.parsedate(row[DEP_IDX]).date()
day = lpputils.parsedate(row[DEP_IDX]).weekday()
hour = lpputils.parsedate(row[DEP_IDX]).hour
minutes = lpputils.parsedate(row[DEP_IDX]).minute
result[day] = 1
if hour < 6:
result[7 + hour] = 1
elif 6 <= hour <= 8:
offset = (hour - 6) * 2
if 0 <= minutes <= 20:
result[7 + hour + offset] = 1
elif 20 < minutes <= 40:
result[7 + hour + offset + 1] = 1
elif 40 < minutes <= 59:
result[7 + hour + offset + 2] = 1
else:
result[7 + hour + 6] = 1
holiday = 0
school_hol = 0
summer_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
if lpputils.parsedate(SUMMER_HOL[0]).date() <= date <= lpputils.parsedate(
SUMMER_HOL[1]).date():
summer_hol = 1
result[-4] = summer_hol
result[-3] = holiday
result[-2] = school_hol
if date.strftime("%Y-%m-%d") in arso.keys():
result[-1] = arso[date.strftime("%Y-%m-%d")][0]
else:
print("No data !!!")
return result
def model10(row):
"""
MODEL10 : binary day and hour attributes + all holiday (binary) -> added 20 min interval between 06 and 09
indeksi : 24 * 7 * 4 kombinacije dan ura, pocitnice 3x, padavine, detour
"""
global arso
result = np.zeros(7 * 24 * 4 + 5)
date = lpputils.parsedate(row[DEP_IDX]).date()
day = lpputils.parsedate(row[DEP_IDX]).weekday()
hour = lpputils.parsedate(row[DEP_IDX]).hour
minutes = lpputils.parsedate(row[DEP_IDX]).minute
day_offset = (24 * 4) * day
if 0 <= minutes <= 15:
result[day_offset + hour * 4] = 1
elif 15 < minutes <= 30:
result[day_offset + hour * 4 + 1] = 1
elif 30 < minutes <= 45:
result[day_offset + hour * 4 + 2] = 1
elif 45 < minutes <= 59:
result[day_offset + hour * 4 + 3] = 1
holiday = 0
school_hol = 0
summer_hol = 0
if date in HOLIDAYS:
holiday = 1
if date in SCHOOL_HOL:
school_hol = 1
if lpputils.parsedate(SUMMER_HOL[0]).date() <= date <= lpputils.parsedate(
SUMMER_HOL[1]).date():
summer_hol = 1
result[-5] = summer_hol
result[-4] = holiday
result[-3] = school_hol
if date.strftime("%Y-%m-%d") in arso.keys():
result[-2] = 1 if arso[date.strftime("%Y-%m-%d")][0] > 20 else 0
else:
print("No data for weather !!!")
line = row[2]
if row[3][0:2] in ['B ', 'G ', 'I ', 'Z ']:
line += row[3][0]
result[-1] = check_detour(line, row[DEP_IDX], detours)
return result
if __name__ == "__main__":
f = gzip.open("train_pred.csv.gz", "rt")
reader = csv.reader(f, delimiter="\t")
next(reader)
# ['Registration', 'Driver ID', 'Route', 'Route Direction', 'Route description', 'First station', 'Departure time', 'Last station', 'Arrival time']
data = [d for d in reader]
noLines = len(data)
Y = numpy.zeros(noLines)
X = numpy.zeros([noLines, 7])
for i, line in enumerate(data):
Y[i] = lpputils.tsdiff(line[-1], line[-3]) # določimo čas vožnje
odhod = lpputils.parsedate(line[-1])
X[i] = getAttributes(odhod)
lr = linear.LinearLearner(lambda_=1.)
napovednik = lr(X, Y)
f = gzip.open("test_pred.csv.gz", "rt")
test = csv.reader(f, delimiter="\t")
next(reader) # skip legend
fo = open("naloga3.txt", "wt")
for l in test:
odhod = lpputils.parsedate(l[-3])
nov_primer = numpy.array(getAttributes(odhod))
#print(nov_primer)
c = napovednik(nov_primer)
else:
linije[primer[3]] = [primer]
linearna_regresija = linear.LinearLearner()
for linija in linije.keys():
x, y = zgradi_matrike(linije[linija], True)
linije[linija] = linearna_regresija(x, y)
f = gzip.open("test.csv.gz", "rt",
encoding="latin1") #za izpis MAE spremeni v "train.csv.gz"
vrstica = csv.reader(f, delimiter="\t")
next(vrstica)
ime, dejanski_cas, primeri, testni_X = zgradi_matrike(vrstica, False)
datoteka = open("napovedi_tekmovanje.txt", "wt", encoding="latin1")
mae_mesec = 11
mae = 0
stevilo_primerov = 0
for vrstica in range(len(primeri)):
napoved = lpputils.tsadd(primeri[vrstica],
-linije[ime[vrstica]](testni_X[vrstica]))
datoteka.write(napoved + "\n")
if lpputils.parsedate(primeri[vrstica]).month == mae_mesec:
mae += absolute_error(dejanski_cas[vrstica], napoved)
stevilo_primerov += 1
datoteka.close()
if mae != 0:
print("Mean absolute error:", mae / stevilo_primerov)
