-
Notifications
You must be signed in to change notification settings - Fork 0
/
Aula7_expressedAgendaModel.R
182 lines (134 loc) · 5.72 KB
/
Aula7_expressedAgendaModel.R
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
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
###################################
#### Manoel Galdino 24/07.2014 ####
#### Aula 7 - Projeto Retórica ####
###################################
#install.packages("RJSONIO")
# pacotes
library("XML")
library("MCMCpack")
library("tm")
library("SnowballC")
library("data.table")
library("ggplot2")
library("seqinr")
library("RTextTools")
library("topicmodels")
library("compiler")
library("RJSONIO")
##
## Para rodar o modelo
## Muda diretório onde estão os dados
setwd("D:\\2014\\aulas\\IESP\\scripts\\Curso-IESP-R-aula\\Data")
load("DTM.RData") # Document Term Matrix
load("Autor_Matrix.RData") # Author Matrix
load("Info.RData")
load("20131030_infoPeqExpArrumado_legis_54.RData")
## script em outro diretório
setwd("D:\\2014\\aulas\\IESP\\scripts\\Curso-IESP-R-aula")
source('ExpAgendVMVA.R')
exp.agenda.vonmon.comp <- cmpfun( exp.agenda.vonmon)
# Originalmente...
# topics <- exp.agenda.vonmon(term.doc = as.matrix(dtm), authors = autorMatrix,
# n.cats = 70,
# verbose = T, kappa = 400)
# Vamos tentar com função compilada
topics <- exp.agenda.vonmon.comp(term.doc = as.matrix(dtm),
authors = autorMatrix,
n.cats = 70,
verbose = T, kappa = 400)
## topics tem o output da func
## precisamos extrair os tópicos
setwd("D:\\2014\\aulas\\IESP\\scripts\\Curso-IESP-R-aula\\Data")
load("20131031_70_topics_peq_Exp_54_legislatura.RData")
load("20131030_topics_peq_Exp_54_legislatura.RData")
# extraiTopicoAutor <- function (topic) {
# autorTopicOne <- NULL
# for( i in 1:dim(topic[[1]])[1]){
# autorTopicOne[i] <- which.max(topic[[1]][i,])
# }
# return(autorTopicOne)
# }
extraiTopicoAutor1 <- function (topic) {
autorTopicOne <- NULL
for( i in 1:dim(topic[[1]][[1]])[1]){
autorTopicOne[i] <- which.max(topic[[1]][[1]][i,])
}
return(autorTopicOne)
}
#autorTopicOne <- extraiTopicoAutor(topics)
autorTopicOne <- extraiTopicoAutor1(topic70)
head(autorTopicOne)
# autorTopicPerc <- prop.table(topics[[1]], 1) # compute the proportion of documents from each author to each topic
autorTopicPerc <- prop.table(topic70[[1]][[1]], 1) # compute the proportion of documents from each author to each topic
head(autorTopicPerc)
autorTopicOne <- as.data.frame(autorTopicOne)
# for( i in 1:nrow(autorTopicOne)){
# autorTopicOne$enfase[i] <- autorTopicPerc[i,which.max(autorTopicPerc[i,])]
# }
#
# json_file <- "https://github.com/Demoulidor/Dados/tree/master/deputadosFederais/deputados.json"
# json_data <- fromJSON(paste(readLines(json_file), collapse=""))
#
# df <- data.frame(nome=json_data[[1]][[1]]$nome,
# url=json_data[[1]][[1]]$url,
# foto=gsub("full\\/", "", json_data[[1]][[1]]$images[[1]]$path),
# email=gsub("mailto:", "", json_data[[1]][[1]]$email),
# id=json_data[[1]][[1]]$id_dep)
#
# for ( i in 2:length(json_data[[1]])) {
# df <- rbind(df, data.frame(nome=json_data[[1]][[i]]$nome,
# url=json_data[[1]][[i]]$url,
# foto=gsub("full\\/", "", json_data[[1]][[i]]$images[[1]]$path),
# email=gsub("mailto:", "", json_data[[1]][[i]]$email),
# id=json_data[[1]][[i]]$id_dep))
# }
autorTopicOne$uf <- infoPeqExpArrumado$uf[!duplicated(infoPeqExpArrumado$autor)]
autorTopicOne$partido <- infoPeqExpArrumado$partido[!duplicated(infoPeqExpArrumado$autor)]
autorTopicOne$autor <- unique(infoPeqExpArrumado$autor)
head(autorTopicOne)
## Qtde de deputados em cada tópico
qplot(autorTopicOne$autorTopicOne, geom="bar", binwidth=1)
table(autorTopicOne$autorTopicOne)
wordTopicPerc <- prop.table(topic70[[1]][[2]], 1) # compute the proportion of documents from each author to each topic
wordTopicPerc[1:5,]
rowSums(wordTopicPerc[1:5,]) # tem q dar 1
StemsTopico <- function(matrix){
# fun??o que recebe uma matriz stem \times topics e retorna um vetor cujo tamanho
# e igual ao numero de topicos e cada elemento representa at? 10 Stems mais
# importantes daquele topico.
wordTopic <- data.frame()
for( i in 1:dim(matrix)[1]){
wordTopic[i, 1] <- which.max(matrix[i,])
wordTopic[i, 2] <- matrix[i, wordTopic[i, 1]]
}
wordTopic$topic <- rownames(matrix)
wordTopic <- wordTopic[order(wordTopic[,1], -wordTopic[,2]), ]
vec <- NULL
for (i in 1:length(unique(wordTopic$V1))){
vec[i] <- paste(subset(wordTopic, V1 == i)[1:10, 3], collapse = " ")
}
# retirando NAs
vec <- gsub("NA", "", vec)
vec <- trimSpace(vec)
vec <- gsub(" ", "; ", vec)
return(vec)
}
wordTopic <- StemsTopico(matrix = topic70[[1]][[2]])
##Analisa as palavras mais importantes do topico
## e determina o conteúdo do tópico
# The goal is to identify words that are common among documents that discuss the same
# topic and rare in documents that were generated by another topic. To identify the
# set of words that satisfy these properties, I select 10 words with the highest mutual information
# with a topic to label the clusters, which provides a principled method for cluster
# labeling appropriate for any unsupervised learning technique.
head(wordTopic)
# autorTopicVis <- merge(autorTopicOne, df, by.x="autor", by.y="nome", all.x=T)
# arrumando sites dos deputados
# autorTopicVis$url <- gsub("http:\\/\\/", "http:\\/\\/www.", autorTopicVis$url)
# rotulando topicos
# rotulos <- read.table('20131031_rotulos1.txt', sep = "\t", header = T)
# autorTopicVis <- merge(autorTopicVis, rotulos[,c('topico','rotulo')], by.x='autorTopicOne', by.y='topico', all.x=T)
# write.table(autorTopicVis , file="20131031_autorTopicVis_70.csv", sep=",", row.names=T)
## Ref
# Paper do Gimmer
# http://web.stanford.edu/~jgrimmer/ExpAgendaFinal.pdf