R: trigram - Revision history

Onnowpurbo at 13:51, 4 November 2018

2018-11-04T13:51:45Z

← Older revision		Revision as of 13:51, 4 November 2018
Line 7:		Line 7:
	library(ggplot2)		library(ggplot2)
	library(ggraph)		library(ggraph)
			library(readtext)

	text <- readtext("out.txt")		text <- readtext("out.txt")

Onnowpurbo at 13:34, 4 November 2018

2018-11-04T13:34:34Z

← Older revision		Revision as of 13:34, 4 November 2018
Line 22:		Line 22:
	count(word1, word2, word3, sort = TRUE)		count(word1, word2, word3, sort = TRUE)
	trigram_graph <- trigram_counts %>%		trigram_graph <- trigram_counts %>%
	filter(n > 10) %>%		filter(n > 20) %>%
	graph_from_data_frame()		graph_from_data_frame()
	trigram_graph		trigram_graph

Onnowpurbo: Created page with " library(dplyr) library(tidytext) library(janeaustenr) library(tidyr) library(igraph) library(ggplot2) library(ggraph) text <- readtext("out.txt") text_trigrams <- t..."

2018-11-04T13:32:40Z

Created page with " library(dplyr) library(tidytext) library(janeaustenr) library(tidyr) library(igraph) library(ggplot2) library(ggraph) text <- readtext("out.txt") text_trigrams <- t..."

New page

library(dplyr)
library(tidytext)
library(janeaustenr)
library(tidyr)
library(igraph)
library(ggplot2)
library(ggraph)

text <- readtext("out.txt")
text_trigrams <- text %>%
unnest_tokens(trigram, text, token = "ngrams", n = 3)
text_trigrams
trigrams_separated <- text_trigrams %>%
separate(trigram, c("word1", "word2", "word3"), sep = " ")
# stopwords Indonesia
trigrams_filtered <- trigrams_separated %>%
filter(!word1 %in% stopwords::stopwords("id", source = "stopwords-iso")) %>%
filter(!word2 %in% stopwords::stopwords("id", source = "stopwords-iso")) %>%
filter(!word3 %in% stopwords::stopwords("id", source = "stopwords-iso"))
trigram_counts <- trigrams_filtered %>%
count(word1, word2, word3, sort = TRUE)
trigram_graph <- trigram_counts %>%
filter(n > 10) %>%
graph_from_data_frame()
trigram_graph
set.seed(2017)
ggraph(trigram_graph, layout = "fr") +
geom_edge_link() +
geom_node_point() +
geom_node_text(aes(label = name), vjust = 1, hjust = 1)