I am trying to use this innovative forecasting tool provided by Facebook called Facebook Prophet (https://research.fb.com/prophet-forecasting-at-scale/). But then I get the following error message when trying to use the "fit" feature. Currently, I use Python on Anaconda environment, with Pystan's 2.14 version, Pandas's 0.20.2 version, and Python's 3.5 version. I don't understand why I encounter this issue, so could anyone please help?
My code:
import pandas as pd
import numpy as np
from fbprophet import Prophet
import matplotlib.pyplot as plt
import pystan
%matplotlib inline
import matplotlib
import matplotlib.pyplot as plt
plt.rcParams['axes.labelsize'] = 14
plt.rcParams['xtick.labelsize'] = 12
plt.rcParams['ytick.labelsize'] = 12
jet=plt.get_cmap('coolwarm')
datapath = "/Users/tuanle/DynamicPricing/Data/"
HAN_BIKE = pd.read_csv(datapath+ "HAN-BIKE_1year_CompletedOrders.csv")
HAN_BIKE['to_char'] = pd.DatetimeIndex(HAN_BIKE['to_char'])
HAN_BIKE = HAN_BIKE.rename(columns={'to_char': 'ds', 'hourr': 'hour','completed_orders': 'y'})
HAN_BIKE['ds'] = HAN_BIKE[['ds','hour']].apply(lambda x : '{}{}'.format(x[0],x[1]), axis=1)
pd.pivot_table(HAN_BIKE, index = ["hour"], values = ["percent_completed", "y"])
HAN_BIKE = HAN_BIKE.drop("hour", axis=1)
HAN_BIKE.hist(bins=50, figsize=(10,15))
HAN_BIKE = HAN_BIKE.drop("percent_completed", axis=1)
# HAN_BIKE.set_index('ds').plot()
HAN_BIKE.loc[(HAN_BIKE['y'] > 500), 'y'] = np.nan
HAN_BIKE.loc[(HAN_BIKE['y'] <= 10), 'y'] = np.nan
HAN_BIKE['y'] = np.log(HAN_BIKE['y'])
# HAN_BIKE.set_index('ds').plot()
HAN_BIKE.hist(bins = 50, figsize = (10,15))
m = Prophet(yearly_seasonality=True)
m.fit(HAN_BIKE)
Error Message:
WARNING:pystan:numpy.core.multiarray failed to import
WARNING:pystan:Something went wrong while unpickling the StanModel. Consider recompiling.
RuntimeError Traceback (most recent call last)
RuntimeError: module compiled against API version 0xb but this version of numpy is 0xa
WARNING:pystan:numpy.core.multiarray failed to import
WARNING:pystan:Something went wrong while unpickling the StanModel. Consider recompiling.
Disabling yearly seasonality. Run prophet with yearly_seasonality=True to override this.
AttributeError Traceback (most recent call last)
in ()
11 HAN_BIKE["ds"] = HAN_BIKE["ds"].values.astype('datetime64[D]')
12 m = Prophet()
---> 13 m.fit(HAN_BIKE)
/Applications/anaconda/lib/python3.5/site-packages/fbprophet/forecaster.py in fit(self, df, **kwargs)
533
534 else:
--> 535 params = model.optimizing(dat, init=stan_init, iter=1e4, **kwargs)
536 for par in params:
537 self.params[par] = params[par].reshape((1, -1))
/Applications/anaconda/lib/python3.5/site-packages/pystan/model.py in optimizing(self, data, seed, init, sample_file, algorithm, verbose, as_vector, **kwargs)
466 data = {}
467
--> 468 fit = self.fit_class(data)
469
470 m_pars = fit._get_param_names()
AttributeError: 'StanModel' object has no attribute 'fit_class''