/
eval_metric_act_agent.py
executable file
·113 lines (91 loc) · 5.03 KB
/
eval_metric_act_agent.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
import pandas as pd
import numpy as np
from sklearn.base import BaseEstimator, TransformerMixin
class Activity_Agent_Evaluation(BaseEstimator,TransformerMixin):
def __init__(self,usage_threshold,activity_threshold,num_devices,num_activities, activity_vector):
self.usage_threshold = usage_threshold
self.activity_threshold = activity_threshold
self.num_devices = num_devices
self.num_activities = num_activities
self.activity_vector = activity_vector
def fit(self,X,y=None):
return self
def transform(self,X,y=None):
data = X.copy()
data = data.tail(24)
data = data.round(3)
l = self.num_devices+self.num_activities
# Get the activity-device mapping
act_vec = pd.read_csv(self.activity_vector, index_col=0, sep = ',')
# Creation of a dictionary with activities as keys and their identifying devices as corresponding values
ident_dev_act={}
for i in range(len(act_vec)):
act = act_vec.index[i]
prob = act_vec.columns[act_vec.loc[act].eq(1)].tolist()
if len(prob)==1:
prob = prob[0]
ident_dev_act.update({act:prob})
ident_dev_act
# Count of total comparisons
total = 0
# Count of right prediction sets
count = 0
for i in range(24):
activity_list = list(data.iloc[i:i+1,self.num_devices:l].columns)
list_dev_keys = []
list_act_prob = []
no_match = False
# Get the activities that would result from the devices with a higher usage probability
# depending on the usage_threshold
for j in range(self.num_devices):
use_prob = data.iloc[i:i+1,:self.num_devices].sort_values(by = data.index[i],axis=1,ascending = False).transpose().iloc[j:j+1,:].values.item()
if use_prob>self.usage_threshold:
dev = data.iloc[i:i+1,:self.num_devices].sort_values(by = data.index[i],axis=1,ascending = False).transpose().iloc[j:j+1,:].index.values
dev = ''.join(map(str,dev)).replace('_usage','')
# activities that are identified by more than one device
# House 5
if dev in (['Tumble_Dryer','Washing_machine']):
dev = ['Tumble_Dryer','Washing_machine']
if dev in (['Microwave','Kettle','Toaster']):
dev = ['Microwave','Kettle','Toaster']
# House 1
#if dev in ['Washer_Dryer', 'Washing_Machine']:
# dev = ['Washer_Dryer', 'Washing_Machine']
# House 2
#if dev in ['Microwave', 'Toaster', 'Kettle']:
# dev = ['Microwave', 'Toaster', 'Kettle']
#if dev in ['Television', 'Hi-Fi']:
# dev = ['Television', 'Hi-Fi']
# House 3
#if dev in ['Tumble_Dryer', 'Washing_Machine']:
# dev = ['Tumble_Dryer', 'Washing_Machine']
#if dev in ['Toaster', 'Microwave', 'Kettle']:
# dev = ['Toaster', 'Microwave', 'Kettle']
# House 4
#if dev in ['Washing_Machine(1)', 'Washing_Machine(2)']:
# dev = ['Washing_Machine(1)', 'Washing_Machine(2)']
#if dev in ['Microwave', 'Kettle']:
# dev = ['Microwave', 'Kettle']
# Get the activity that uses the devices as identifying one(s)
act = [k for k, v in ident_dev_act.items() if v == dev][0]
if act in activity_list:
list_dev_keys.append(act)
activity_list.remove(act)
# Get the activities the Activity Agent predicted
# depending on the activity_threshold
for j in range(self.num_activities):
act_prob = data.iloc[i:i+1,l:].sort_values(by = data.index[i],axis=1,ascending = False).transpose().iloc[j:j+1,:].values.item()
if act_prob>self.activity_threshold:
act_prob = data.iloc[i:i+1,l:].sort_values(by = data.index[i],axis=1,ascending = False).transpose().iloc[j:j+1,:].index.values
act_prob = ''.join(map(str,act_prob)).replace('_prob','')
list_act_prob.append(act_prob)
# Check if both lists contain equal elements
check = any(item in list_dev_keys for item in list_act_prob) #all
if check is True and len(list_dev_keys) and len(list_act_prob):
count+=1
total+=1
else:
total+=1
# Compute EQUAL score (ratio of equal sets to total number of sets)
EQUAL = count/total
return EQUAL