Bayesian-Comment-Volume-Prediction/construct_graph.py at master · PavanGJ/Bayesian-Comment-Volume-Prediction · GitHub

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
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
from __future__ import division
#############################################################################################################
#
#
#
#	Date 				Name	 		Description
#	24-Mar-2017			Anurag Dixit		Added the file for creating the data for hybridBayesianNetwork Model
#	24-Mar-2017			Anurag Dixit		Bug fix (mean_scale for Comments)
#	24-Mar-2017			Pavan Joshi		Minor Bug fix to overcome indexing problems in some systems
#	25-Mar-2017			Pavan Joshi		Minor fixes to overcome bugs when serializing numpy array and values
#	26-Mar-2017			Pavan Joshi		Bug fixes for creation of dictionary that affected further processing
#	26-Mar-2017			Pavan Joshi		Added functionality to handle continuous parents for hybrid nodes
#
############################################################################################################

import numpy as np
from sklearn import linear_model
from Data import Data

class Ndata(Data):

	def __init__(self, fname):

		Data.__init__(self, fname)


	def linear_regr(self, nodeValue, parents):

		reg = linear_model.LinearRegression()
		reg.fit(parents, nodeValue)
		return reg.coef_, reg.intercept_


	def get_value(self, nodeName, val_type):
		print "get_value", nodeName, val_type

		st = self.model #Structure()
		var_index = self.dictIdx[nodeName]
		node_val = self.dictVal[nodeName]
		parents = st.get_parents(nodeName)
		children = st.get_children(nodeName)


		if(val_type == "lgandd"):

			discrete_parents = []
			non_discrete = []
			for i in parents:
				#print i,"type",st.get_vertex_type(i)
				if(st.get_vertex_type(i) == "d"):
					discrete_parents.append(i)
				else:
					non_discrete.append(i)


			print "Discrete parents", discrete_parents
			disc_parents = self.dictVal[discrete_parents[0]]
			disc_index = self.dictIdx[discrete_parents[0]]
			values = np.array(np.unique(disc_parents),  dtype = 'float')

			count = len(values)

			vals = []
			variance = []
			mean_base = []
			mean_scale = []


			for i in values:
				filter_val = self.data[:, var_index][np.array(self.data[:, disc_index],dtype=np.float) == i]

				a = np.array(filter_val).astype(np.float)
				if len(non_discrete)==0:
					mean_base.append(float(np.mean(a)))
					mean_scale.append([])
				else:
					v = self.dictVal[non_discrete[0]]
					v = v[np.array(self.data[:, disc_index],dtype=np.float) == i]
					parent = v.reshape(len(v),1)

					for i in range(1, len(non_discrete)):
						val = self.dictVal[non_discrete[i]][np.array(self.data[:, disc_index],dtype=np.float) == i]
						parent = np.concatenate((parent, val.reshape(len(val), 1)), axis=1).astype(np.float)

					node_val = filter_val.reshape(len(filter_val), 1).astype(np.float)
					w, beta_0 = self.linear_regr(node_val, parent)

					mean_base.append(beta_0.tolist()[0])
					mean_scale.append([y for x in w.tolist() for y in x])
				vals.append("['%s']" % float(i))
				variance.append(float(np.var(a)))


			d = {}
			for i in range(0, len(vals)):

				dic = {vals[i] : {"mean_base":mean_base[i], "variance": variance[i], "mean_scal":mean_scale[i]} }
				d.update(dic)


			ret = {"type":val_type, "parents":parents, "children":children, "hybcprob": d}

			return ret

		elif(val_type == "d"):

			values = np.unique(node_val)
			count = len(values)
			total = len(self.data[:,0])

			prob = []
			for i in values:
				filter_val = np.array(self.data[np.array(self.data[:, var_index],dtype=np.float)==float(i)])
				prob.append(len(filter_val)/total)

			ret = {"type": "discrete","parents": parents, "children": children, "numoutcomes": float(count), "cprob": [float(x) for x in prob], "vals":[float(x) for x in values]}
			return ret

		else:   #linear gaussian types
			mean_scal = []
			mean_base = None

			node_val = node_val.reshape(len(node_val), 1).astype(np.float)
			variance = float(np.var(node_val))
			if(parents == None):
				mean_base = np.mean(node_val)
				#mean_scal.append(1)


			else:

				v = self.dictVal[parents[0]]
				parent = v.reshape(len(v),1)

				for i in range(1, len(parents)):
					val = self.dictVal[parents[i]]
					parent = np.concatenate((parent, val.reshape(len(val), 1)), axis=1).astype(np.float)


				w, beta_0 = self.linear_regr(node_val, parent)

				mean_base = beta_0.tolist()[0]
				mean_scal = [y for x in w.tolist() for y in x]

			ret = {"type":val_type, "parents":parents, "children":children, "mean_base": mean_base, "mean_scal":[x for x in mean_scal], "variance": variance}

			return ret