00db6a30ad
- Refactoring generate_flrg and train methods - Introducing batches and model saving on fit method
105 lines
3.3 KiB
Python
105 lines
3.3 KiB
Python
"""
|
||
First Order Exponentialy Weighted Fuzzy Time Series by Sadaei et al. (2013)
|
||
|
||
H. J. Sadaei, R. Enayatifar, A. H. Abdullah, and A. Gani, “Short-term load forecasting using a hybrid model with a
|
||
refined exponentially weighted fuzzy time series and an improved harmony search,” Int. J. Electr. Power Energy Syst., vol. 62, no. from 2005, pp. 118–129, 2014.
|
||
"""
|
||
|
||
import numpy as np
|
||
from pyFTS.common import FuzzySet,FLR,fts, flrg
|
||
|
||
default_c = 1.1
|
||
|
||
|
||
class ExponentialyWeightedFLRG(flrg.FLRG):
|
||
"""First Order Exponentialy Weighted Fuzzy Logical Relationship Group"""
|
||
def __init__(self, LHS, **kwargs):
|
||
super(ExponentialyWeightedFLRG, self).__init__(1, **kwargs)
|
||
self.LHS = LHS
|
||
self.RHS = []
|
||
self.count = 0.0
|
||
self.c = kwargs.get("c",default_c)
|
||
self.w = None
|
||
|
||
def append(self, c):
|
||
self.RHS.append(c)
|
||
self.count = self.count + 1.0
|
||
|
||
def weights(self):
|
||
if self.w is None:
|
||
wei = [self.c ** k for k in np.arange(0.0, self.count, 1.0)]
|
||
tot = sum(wei)
|
||
self.w = np.array([k / tot for k in wei])
|
||
return self.w
|
||
|
||
def __str__(self):
|
||
tmp = self.LHS.name + " -> "
|
||
tmp2 = ""
|
||
cc = 0
|
||
wei = [self.c ** k for k in np.arange(0.0, self.count, 1.0)]
|
||
tot = sum(wei)
|
||
for c in sorted(self.RHS, key=lambda s: s.name):
|
||
if len(tmp2) > 0:
|
||
tmp2 = tmp2 + ","
|
||
tmp2 = tmp2 + c.name + "(" + str(wei[cc] / tot) + ")"
|
||
cc = cc + 1
|
||
return tmp + tmp2
|
||
|
||
def __len__(self):
|
||
return len(self.RHS)
|
||
|
||
|
||
class ExponentialyWeightedFTS(fts.FTS):
|
||
"""First Order Exponentialy Weighted Fuzzy Time Series"""
|
||
def __init__(self, name, **kwargs):
|
||
super(ExponentialyWeightedFTS, self).__init__(1, "EWFTS", **kwargs)
|
||
self.name = "Exponentialy Weighted FTS"
|
||
self.detail = "Sadaei"
|
||
self.c = kwargs.get('c', default_c)
|
||
|
||
def generate_flrg(self, flrs, c):
|
||
for flr in flrs:
|
||
if flr.LHS.name in self.flrgs:
|
||
self.flrgs[flr.LHS.name].append(flr.RHS)
|
||
else:
|
||
self.flrgs[flr.LHS.name] = ExponentialyWeightedFLRG(flr.LHS, c=c);
|
||
self.flrgs[flr.LHS.name].append(flr.RHS)
|
||
|
||
def train(self, data, **kwargs):
|
||
self.c = kwargs.get('parameters', default_c)
|
||
if kwargs.get('sets', None) is not None:
|
||
self.sets = kwargs.get('sets', None)
|
||
ndata = self.apply_transformations(data)
|
||
tmpdata = FuzzySet.fuzzyfy_series_old(ndata, self.sets)
|
||
flrs = FLR.generate_recurrent_flrs(tmpdata)
|
||
self.generate_flrg(flrs, self.c)
|
||
|
||
def forecast(self, data, **kwargs):
|
||
l = 1
|
||
|
||
data = np.array(data)
|
||
|
||
ndata = self.apply_transformations(data)
|
||
|
||
l = len(ndata)
|
||
|
||
ret = []
|
||
|
||
for k in np.arange(0, l):
|
||
|
||
mv = FuzzySet.fuzzyfy_instance(ndata[k], self.sets)
|
||
|
||
actual = self.sets[np.argwhere(mv == max(mv))[0, 0]]
|
||
|
||
if actual.name not in self.flrgs:
|
||
ret.append(actual.centroid)
|
||
else:
|
||
flrg = self.flrgs[actual.name]
|
||
mp = flrg.get_midpoints()
|
||
|
||
ret.append(mp.dot(flrg.weights()))
|
||
|
||
ret = self.apply_inverse_transformations(ret, params=[data])
|
||
|
||
return ret
|