-
Notifications
You must be signed in to change notification settings - Fork 0
/
liwc_regression.R
81 lines (70 loc) · 3.69 KB
/
liwc_regression.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
# library(leaps)
library("writexl")
# training_data <- head(liwc_filtered, 536)
# liwc_model <- lm(Q5Mean ~ ppron + ipron + article + prep + shehe + focusfuture + focuspresent + relativ + informal + WPS + auxverb + adverb + verb + adj +
# compare + posemo + negemo + affect + quant + affiliation + achieve + reward + insight + cause + discrep + certain + tentat
# + percept + bio + work + money + relig + death + family, data=liwc_filtered)
#
# adjusted_model <- lm(Q5Mean ~ ipron + adverb + money + verb + cogproc + article + focusfuture + affiliation, data=training_data)
#
#summary(liwc_model)
#summary(adjusted_model)
#
# models <- regsubsets(Q5Mean ~ ppron + ipron + article + prep + shehe + focusfuture + focuspresent + relativ + informal + WPS + auxverb + adverb + verb + adj +
# compare + posemo + negemo + affect + quant + affiliation + achieve + reward + insight + cause + discrep + certain + tentat
# + percept + bio + work + money + relig + death + family, nvmax = 10,data=liwc_filtered)
# adjusted_subsets <- regsubsets(Q5Mean ~ adverb + verb + adj + affiliation + insight + cause + money + focusfuture + ipron + article, nvmax = 10, data=liwc_filtered)
#
# summary(adjusted_subsets)
#
training_data$log_mean <- log(as.numeric(training_data$Q5Mean))
training_data <- training_data[-c(375),]
# hist(log(training_data$auxverb))
# experiment_model <- lm(log_mean ~ ., data = training_data)
# summary(experiment_model)
# plot(experiment_model, which=1)
# experiment_subsets <- regsubsets(log_mean ~
# pronoun + ppron + ipron + auxverb + `function` + number + sad + social + male + percept + see + hear + feel + affiliation + focusfuture + time + work + money + relig + informal + swear,
# data = training_data, nvmax = 15)
# summary(experiment_subsets)
#
# plot(log_mean ~ auxverb, data=training_data)
final_model <- lm(log_mean ~
male + focusfuture + time + work + money + relig,
data=training_data)
# total_model <- lm(log_mean ~ ., data=training_data)
summary(final_model)
# summary(experiment_model)
# summary(total_model)
# plot(training_data$female, training_data$log_mean)
plot(final_model)
# log.data <- training_data
# log.data$mean <- as.numeric(head(liwc_filtered, 536)['Q5Mean'][-c(375),])
# log.data <- log(log.data)
# is.na(log.data)<-sapply(log.data, is.infinite)
# log.data[is.na(log.data)]<-0
# log.experiment_model = lm(mean ~
# .,
# data=log.data)
# summary(log.experiment_model)
# plot(log.data$bio, log.data$mean)
# (testing_data <-> testing_data2) (final_model <-> total_model)
testing_data <- tail(liwc_filtered, 536)
testing_data$log_mean <- log(as.numeric(testing_data$Q5Mean))
testing_data$prediction <- predict(final_model, testing_data, interval="prediction")
testing_data$difference <- testing_data$prediction - testing_data$log_mean
summary(testing_data$difference)
testing_data$adjusted_difference <- exp(testing_data$difference)
summary(testing_data$adjusted_difference)
testing_data$actual_prediction <- exp(testing_data$prediction)
summary(testing_data$actual_prediction)
testing_data$actual_difference <- testing_data$actual_prediction - as.numeric(testing_data$Q5Mean)
summary(abs(testing_data$actual_difference))
testing_data$Q5Mean <- as.numeric(testing_data$Q5Mean)
summary(as.numeric(testing_data$Q5Mean))
summary(as.numeric(testing_data$actual_prediction))
summary(as.numeric(abs(testing_data$actual_difference)))
plot(testing_data$male, log(as.numeric(testing_data$Q5Mean)))
##
cor.test(training_data[["log_mean"]], training_data[["WC"]])
write_xlsx(training_data,"DATA\\trainingdata.xlsx")