MNAnalytics/gamelogs.R at main · jtkulas/MNAnalytics · 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
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
181
182
183
184
185
ind.games <- read.csv("MDataFiles_Stage1//MRegularSeasonCompactResults.csv")
teams <- read.csv("MDataFiles_Stage1//MTeams.csv")
teams <- teams[,c(1:2)]
## Scrub out 2009 and earlier (predictors only 2010 through 2021)

twenty10 <- read.csv("2010.csv")
twenty11 <- read.csv("2011.csv")
twenty12 <- read.csv("2012.csv")
twenty13 <- read.csv("2013.csv")
twenty14 <- read.csv("2014.csv")
twenty15 <- read.csv("2015.csv")
twenty16 <- read.csv("2016.csv")
twenty17 <- read.csv("2017.csv")
twenty18 <- read.csv("2018.csv")
twenty19 <- read.csv("2019.csv")

current <- read.csv("current.csv")

write.csv(teams$TeamName, "names_games.csv")
write.csv(twenty10$X.1, "names_predict.csv")

twenty10$X.1 <- gsub("NCAA", "", twenty10$X.1)   ## predictors have "NCAA" next to tournament teams - prob still have match issues but will be better
twenty11$X.1 <- gsub("NCAA", "", twenty11$X.1)
twenty12$X.1 <- gsub("NCAA", "", twenty12$X.1)
twenty13$X.1 <- gsub("NCAA", "", twenty13$X.1)
twenty14$X.1 <- gsub("NCAA", "", twenty14$X.1)
twenty15$X.1 <- gsub("NCAA", "", twenty15$X.1)
twenty16$X.1 <- gsub("NCAA", "", twenty16$X.1)
twenty17$X.1 <- gsub("NCAA", "", twenty17$X.1)
twenty18$X.1 <- gsub("NCAA", "", twenty18$X.1)
twenty19$X.1 <- gsub("NCAA", "", twenty19$X.1)
current$X.1 <- gsub("NCAA", "", current$X.1)

twenty10 <- twenty10[-1,]
twenty11 <- twenty11[-1,]
twenty12 <- twenty12[-1,]
twenty13 <- twenty13[-1,]
twenty14 <- twenty14[-1,]
twenty15 <- twenty15[-1,]
twenty16 <- twenty16[-1,]
twenty17 <- twenty17[-1,]
twenty18 <- twenty18[-1,]
twenty19 <- twenty19[-1,]
current <- current[-1,]

use <- subset(ind.games, Season > 2009)
together <- merge(ind.games, teams, by.x = "WTeamID", by.y = "TeamID")
both <- merge(together, teams, by.x = "LTeamID", by.y = "TeamID")

names(both)[names(both) == "TeamName.x"] <- "Winner"
names(both)[names(both) == "TeamName.y"] <- "Loser"

game2010 <- subset(both, Season == 2010)
game2011 <- subset(both, Season == 2011)
game2012 <- subset(both, Season == 2012)
game2013 <- subset(both, Season == 2013)
game2014 <- subset(both, Season == 2014)
game2015 <- subset(both, Season == 2015)
game2016 <- subset(both, Season == 2016)
game2017 <- subset(both, Season == 2017)
game2018 <- subset(both, Season == 2018)
game2019 <- subset(both, Season == 2019)

to_2010 <- merge(game2010, twenty10, by.x = "Winner", by.y = "X.1")   ## Adding winning team predictors
to_2012 <- merge(game2012, twenty12, by.x = "Winner", by.y = "X.1")
to_2014 <- merge(game2014, twenty14, by.x = "Winner", by.y = "X.1")
to_2016 <- merge(game2016, twenty16, by.x = "Winner", by.y = "X.1")
to_2018 <- merge(game2018, twenty18, by.x = "Winner", by.y = "X.1")


bo_2010 <- merge(to_2010, twenty10, by.x = "Loser", by.y = "X.1")     ## Adding losing team predictors
bo_2012 <- merge(to_2012, twenty12, by.x = "Loser", by.y = "X.1")
bo_2014 <- merge(to_2014, twenty14, by.x = "Loser", by.y = "X.1")
bo_2016 <- merge(to_2016, twenty16, by.x = "Loser", by.y = "X.1")
bo_2018 <- merge(to_2018, twenty18, by.x = "Loser", by.y = "X.1")

big_predictor_win <- rbind(bo_2010, bo_2012, bo_2014, bo_2016, bo_2018)

big_predictor_win$DV1 <- 1
big_predictor_win$DV3 <- big_predictor_win$WScore - big_predictor_win$LScore

## need to duplicate for loser but swap winner and losers

toq_2011 <- merge(game2011, twenty11, by.x = "Loser", by.y = "X.1")
toq_2013 <- merge(game2013, twenty13, by.x = "Loser", by.y = "X.1")
toq_2015 <- merge(game2015, twenty15, by.x = "Loser", by.y = "X.1")
toq_2017 <- merge(game2017, twenty17, by.x = "Loser", by.y = "X.1")
toq_2019 <- merge(game2019, twenty19, by.x = "Loser", by.y = "X.1")

boq_2011 <- merge(toq_2011, twenty11, by.x = "Winner", by.y = "X.1")
boq_2013 <- merge(toq_2013, twenty13, by.x = "Winner", by.y = "X.1")
boq_2015 <- merge(toq_2015, twenty15, by.x = "Winner", by.y = "X.1")
boq_2017 <- merge(toq_2017, twenty17, by.x = "Winner", by.y = "X.1")
boq_2019 <- merge(toq_2019, twenty19, by.x = "Winner", by.y = "X.1")

big_predictor_lose <- rbind(boq_2011, boq_2013, boq_2015, boq_2017, boq_2019)

big_predictor_lose$DV1 <- 0
big_predictor_lose$DV3 <- big_predictor_lose$LScore - big_predictor_lose$WScore

tlose <- t(big_predictor_lose)
tsort <- tlose[order(row.names(tlose)),]
big_predictor_lose <- as.data.frame(t(tsort))

twin <- t(big_predictor_win)
tsortw <- twin[order(row.names(twin)),]
big_predictor_win <- as.data.frame(t(tsortw))

finally <- rbind(big_predictor_win, big_predictor_lose)

library(caret)

control <- trainControl(method="cv", number=10)
metric <- "RMSE"

## psych::describe(finally)
finally <- finally[,-c(27,28,31,32,35,36,71,72,75,76)] ## data not completely good but tourney starts manana
finally2 <- finally[,c(7,14:19,25:68)]
dat2 <- data.frame(lapply(finally2, function(x) as.numeric(as.character(x))))

user <- dat2[complete.cases(dat2), ]

# user$DV2 <- as.factor(as.character(user$DV1))

memory.limit(16500)
fit.lm <- train(DV3 ~ ., data=user, method="lm", metric=metric, trControl=control)
# summary(fit.lm)

## write.csv(game2010, "template.csv")


current <- read.csv("current_bracket.csv")
current <- current[-1,]

both2 <- read.csv("bracket.csv")

pred_to_2010 <- merge(both2, current, by.x = "Winner", by.y = "X.1")
pred_bo_2010 <- merge(pred_to_2010, current, by.x = "Loser", by.y = "X.1")

tlose <- t(pred_bo_2010)
tsort <- tlose[order(row.names(tlose)),]
big_validate <- as.data.frame(t(tsort))

psych::describe(big_validate)

finally3 <- big_validate[,c(12:17,23:24,27:28,31:32,35:68,71:72,75:76)]
validate <- data.frame(lapply(finally3, function(x) as.numeric(as.character(x))))

predictions <- predict(fit.lm, validate)
write.csv(predictions, "winnerR1.csv")


####
####
#### Logistic regression

## Need to swap DV3 with DV1

finally <- rbind(big_predictor_win, big_predictor_lose)
finally <- finally[,-c(27,28,31,32,35,36,71,72,75,76)] ## data not completely good but tourney starts manana
finally2 <- finally[,c(6,14:19,25:68)]
dat2 <- data.frame(lapply(finally2, function(x) as.numeric(as.character(x))))

user <- dat2[complete.cases(dat2), ]


trControl2 <- trainControl(method = 'repeatedcv',
                          number = 5,
                          repeats =  5,
                          search = 'random')

logit.CV <- train(DV1 ~ ., data=user,
                  method = 'glmnet',
                  trControl = trControl2,
                  family = 'binomial' )


user$DV1 <- as.factor(as.character(user$DV1))
logit.CV2 <- train(DV1 ~ ., data=user,
                  method = 'glmnet',
                  trControl = trControl2,
                  family = 'binomial' )

predictions2 <- predict(logit.CV2, validate)
write.csv(predictions2, "winnerLOGISTIC.csv")