Module fcmpy.simulator.inference
Expand source code
import numpy as np
from abc import ABC, abstractmethod
class Inference(ABC):
"""
Class of FCM inference methods
"""
@abstractmethod
def infer() -> np.array:
raise NotImplementedError('Infer method is not defined!')
class Kosko(Inference):
"""
Kosko's inference method.
"""
@staticmethod
def infer(**kwargs) -> np.array:
"""
Kosko's inference method.
Parameters
----------
initial_state: numpy.array
initial state vector of the concepts
weight_matrix: numpy.ndarray
N*N weight matrix of the FCM.
Return
-------
y: numpy.array
updated state vector
"""
initial_state = kwargs['initial_state']
weight_matrix = kwargs['weight_matrix']
weight_matrix = weight_matrix.T
res = weight_matrix.dot(initial_state)
return res
class ModifiedKosko(Inference):
"""
Modified Kosko inference method.
"""
@staticmethod
def infer(**kwargs) -> np.array:
"""
Modified Kosko inference method.
Parameters
----------
initial_state: numpy.array
initial state vector of the concepts
weight_matrix: numpy.ndarray
N*N weight matrix of the FCM.
Return
----------
y: numpy.array
updated state vector
"""
initial_state = kwargs['initial_state']
weight_matrix = kwargs['weight_matrix']
weight_matrix = weight_matrix.T
res = weight_matrix.dot(initial_state) + initial_state
return res
class Rescaled(Inference):
"""
Rescaled inference method.
"""
@staticmethod
def infer(**kwargs) -> np.array:
"""
Rescaled inference method.
Parameters
----------
initial_state: numpy.array
initial state vector of the concepts
weight_matrix: numpy.ndarray
N*N weight matrix of the FCM.
Return
----------
y: numpy.array
updated state vector
"""
initial_state = kwargs['initial_state']
weight_matrix = kwargs['weight_matrix']
weight_matrix = weight_matrix.T
res = weight_matrix.dot(([2*i-1 for i in initial_state])) + ([2*i-1 for i in initial_state])
return resClasses
class Inference-
Class of FCM inference methods
Expand source code
class Inference(ABC): """ Class of FCM inference methods """ @abstractmethod def infer() -> np.array: raise NotImplementedError('Infer method is not defined!')Ancestors
- abc.ABC
Subclasses
Methods
def infer() ‑>-
Expand source code
@abstractmethod def infer() -> np.array: raise NotImplementedError('Infer method is not defined!')
class Kosko-
Kosko's inference method.
Expand source code
class Kosko(Inference): """ Kosko's inference method. """ @staticmethod def infer(**kwargs) -> np.array: """ Kosko's inference method. Parameters ---------- initial_state: numpy.array initial state vector of the concepts weight_matrix: numpy.ndarray N*N weight matrix of the FCM. Return ------- y: numpy.array updated state vector """ initial_state = kwargs['initial_state'] weight_matrix = kwargs['weight_matrix'] weight_matrix = weight_matrix.T res = weight_matrix.dot(initial_state) return resAncestors
- Inference
- abc.ABC
Static methods
def infer(**kwargs) ‑>-
Kosko's inference method.
Parameters
initial_state:numpy.array- initial state vector of the concepts
weight_matrix:numpy.ndarray- N*N weight matrix of the FCM.
Return
y: numpy.array updated state vector
Expand source code
@staticmethod def infer(**kwargs) -> np.array: """ Kosko's inference method. Parameters ---------- initial_state: numpy.array initial state vector of the concepts weight_matrix: numpy.ndarray N*N weight matrix of the FCM. Return ------- y: numpy.array updated state vector """ initial_state = kwargs['initial_state'] weight_matrix = kwargs['weight_matrix'] weight_matrix = weight_matrix.T res = weight_matrix.dot(initial_state) return res
class ModifiedKosko-
Modified Kosko inference method.
Expand source code
class ModifiedKosko(Inference): """ Modified Kosko inference method. """ @staticmethod def infer(**kwargs) -> np.array: """ Modified Kosko inference method. Parameters ---------- initial_state: numpy.array initial state vector of the concepts weight_matrix: numpy.ndarray N*N weight matrix of the FCM. Return ---------- y: numpy.array updated state vector """ initial_state = kwargs['initial_state'] weight_matrix = kwargs['weight_matrix'] weight_matrix = weight_matrix.T res = weight_matrix.dot(initial_state) + initial_state return resAncestors
- Inference
- abc.ABC
Static methods
def infer(**kwargs) ‑>-
Modified Kosko inference method.
Parameters
initial_state:numpy.array- initial state vector of the concepts
weight_matrix:numpy.ndarray- N*N weight matrix of the FCM.
Return
y: numpy.array updated state vector
Expand source code
@staticmethod def infer(**kwargs) -> np.array: """ Modified Kosko inference method. Parameters ---------- initial_state: numpy.array initial state vector of the concepts weight_matrix: numpy.ndarray N*N weight matrix of the FCM. Return ---------- y: numpy.array updated state vector """ initial_state = kwargs['initial_state'] weight_matrix = kwargs['weight_matrix'] weight_matrix = weight_matrix.T res = weight_matrix.dot(initial_state) + initial_state return res
class Rescaled-
Rescaled inference method.
Expand source code
class Rescaled(Inference): """ Rescaled inference method. """ @staticmethod def infer(**kwargs) -> np.array: """ Rescaled inference method. Parameters ---------- initial_state: numpy.array initial state vector of the concepts weight_matrix: numpy.ndarray N*N weight matrix of the FCM. Return ---------- y: numpy.array updated state vector """ initial_state = kwargs['initial_state'] weight_matrix = kwargs['weight_matrix'] weight_matrix = weight_matrix.T res = weight_matrix.dot(([2*i-1 for i in initial_state])) + ([2*i-1 for i in initial_state]) return resAncestors
- Inference
- abc.ABC
Static methods
def infer(**kwargs) ‑>-
Rescaled inference method.
Parameters
initial_state:numpy.array- initial state vector of the concepts
weight_matrix:numpy.ndarray- N*N weight matrix of the FCM.
Return
y: numpy.array updated state vector
Expand source code
@staticmethod def infer(**kwargs) -> np.array: """ Rescaled inference method. Parameters ---------- initial_state: numpy.array initial state vector of the concepts weight_matrix: numpy.ndarray N*N weight matrix of the FCM. Return ---------- y: numpy.array updated state vector """ initial_state = kwargs['initial_state'] weight_matrix = kwargs['weight_matrix'] weight_matrix = weight_matrix.T res = weight_matrix.dot(([2*i-1 for i in initial_state])) + ([2*i-1 for i in initial_state]) return res