def predictions(dataframe,models):
"""we assume model has been trained;
we also assume that the mapper has been under modifys"""
mods = dataframe['module_category'].unique()
dataframe['predictions']=5
for each_mod in mods:
cats = dataframe[dataframe['module_category']==each_mod]['component_category'].unique()
for each_cat in cats:
key =(each_mod,each_cat)
if key in models[0]:
A,mu,sigma = models[0][key]
#A*np.exp(-(x-mu)**2/(2.*sigma**2))
dataframe[(dataframe['module_category']==each_mod) \
& (dataframe['component_category']==each_cat)]['predictions'] = \
A*np.exp(-(dataframe[(dataframe['module_category']==each_mod) \
& (dataframe['component_category']==each_cat)]['time'] - mu)**2/(2.*sigma**2))
dataframe['predictions'].round()
return dataframe
def gauss(x, *p):
A, mu, sigma = p
return A*np.exp(-(x-mu)**2/(2.*sigma**2))
