Topic Modeling

library(topicmodels)
library(tm)

data("AssociatedPress")

inspect(AssociatedPress[1:5, 1:5])

<<DocumentTermMatrix (documents: 5, terms: 5)>>
Non-/sparse entries: 0/25
Sparsity           : 100%
Maximal term length: 10
Weighting          : term frequency (tf)
Sample             :
      Terms
Docs   aaron abandon abandoned abandoning abbott
  [1,]     0       0         0          0      0
  [2,]     0       0         0          0      0
  [3,]     0       0         0          0      0
  [4,]     0       0         0          0      0
  [5,]     0       0         0          0      0

AP_topic_model<-LDA(AssociatedPress, 
                    k=10, 
                    control = list(seed = 321))

library(tidytext)
library(dplyr)

AP_topics <- tidy(AP_topic_model, matrix = "beta")

ap_top_terms <- 
  AP_topics %>%
    group_by(topic) %>%
      top_n(10, beta) %>%
        ungroup() %>%
          arrange(topic, -beta)

library(ggplot2)

ap_top_terms %>%
  mutate(term = reorder(term, beta)) %>%
    mutate(topic = paste("Topic #", topic)) %>%
    ggplot(aes(term, beta, fill = factor(topic))) +
      geom_col(show.legend = FALSE) +
        facet_wrap(~ topic, scales = "free") +
          theme_minimal()+
          theme(plot.title = 
            element_text(hjust = 0.5, size=18))+
          labs(
            title = "Topic Model of AP News Articles",
            caption = "Top Terms by Topic (betas)"
          )+
          ylab("")+
          xlab("")+
          coord_flip()

google_doc_id <- "1LcX-JnpGB0lU1iDnXnxB6WFqBywUKpew" 

poliblogs<-read.csv(
  sprintf("https://docs.google.com/uc?id=%s&export=download", 
          google_doc_id), stringsAsFactors = FALSE)

library(stm)
processed <- textProcessor(poliblogs$documents, 
                           metadata = poliblogs)

out <- prepDocuments(processed$documents, 
                     processed$vocab, 
                     processed$meta)

docs <- out$documents
vocab <- out$vocab
meta <-out$meta

First_STM <- stm(documents = out$documents, vocab = out$vocab,
              K = 10, prevalence =~ rating + s(day) ,
              max.em.its = 75, data = out$meta,
              init.type = "Spectral", verbose = FALSE)

plot(First_STM)

findThoughts(First_STM, texts = poliblogs$documents,
     n = 2, topics = 3)

findingk <- searchK(out$documents, out$vocab, K = c(10, 30),
 prevalence =~ rating + s(day), data = meta, verbose=FALSE)

plot(findingk)

predict_topics<-estimateEffect(formula = 1:10 ~
                                rating + 
                                s(day), 
                                stmobj = First_STM, 
                                metadata = out$meta, 
                                uncertainty = "Global")

plot(predict_topics, covariate = "rating", 
     topics = c(3, 5, 9), model = First_STM, 
     method = "difference",
     cov.value1 = "Liberal", cov.value2 = "Conservative",
     xlab = "More Conservative ... More Liberal",
     main = "Effect of Liberal vs. Conservative",
     xlim = c(-.1, .1), labeltype = "custom",
     custom.labels = c('Topic 3', 'Topic 5','Topic 9'))

plot(predict_topics, "day", method = "continuous", 
     topics = 3, model = z, printlegend = FALSE, 
     xaxt = "n", xlab = "Time (2008)")
     monthseq <- seq(from = as.Date("2008-01-01"),
     to = as.Date("2008-12-01"), by = "month")
     monthnames <- months(monthseq)
     axis(1,at = as.numeric(monthseq) - 
     min(as.numeric(monthseq)), labels = monthnames)