dota_mcts.Rmd

---
title: "dota_mcts"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```


```{r}
setwd('/Users/Isolachine/Downloads')
df <- read.csv('lineups.csv')
hero <- read.csv('hero_data.csv')
```

```{r}
summary(df)
dim(df)[1]
```

```{r}
df.all <- data.frame()
df.sparse <- data.frame()

processed = 0

for (i in 1:dim(df)[1]) {
  game.vector <- matrix(data = 0, ncol = 113)
  count = 0
  for (v in df[i,]) {
    if (count == 0) {
      game.vector[113] = as.integer(v) - 1
    } else if (count <= 5) {
      game.vector[v] <- 1
    } else {
      game.vector[v] <- -1
    }
    count = count + 1
  }
  names(game.vector) <- 1:113
  df.sparse <- rbind.data.frame(df.sparse, game.vector)
  processed = processed + 1
  if (processed %% 1000 == 0) {
    print(paste('processed:', processed))
    df.all <- rbind.data.frame(df.all, df.sparse)
    df.sparse <- data.frame()
  }
}
```

```{r}
df.sparse[,113]
```

```{r}
write.csv(df.all, file = 'sparse.csv')
```


```{r}
cutoff <- 30000
trainset <- df.all[1:cutoff, ]
testset <- df.all[(cutoff + 1):dim(df.all)[1], ]
```


```{r}
n <- names(df.all)
formula <- as.formula(paste("V113 ~", paste(n[!n %in% "V113"], collapse = " + ")))
library(neuralnet)
```


```{r}
nn <- neuralnet(formula = formula, data = trainset, hidden=0, linear.output=FALSE, threshold=0.01)
nn$result.matrix
plot(nn)
```

```{r}
#Test the resulting output
temp_test <- subset(testset, select = n[!n %in% "V113"])
head(temp_test)
nn.results <- compute(nn, temp_test)
results <- data.frame(actual = testset$V113, prediction = nn.results$net.result)
results
```

```{r}
roundedresults<-sapply(results,round,digits=0)
roundedresultsdf=data.frame(roundedresults)
attach(roundedresultsdf)
cf <- table(actual,prediction)
cf

acc <- (cf[1,1] + cf[2,2]) / sum(cf)
acc

```