ratings.train = read.csv("/Users/fay_kong/Desktop/SML_SPRING/FILM/movies_matchup.csv", head = TRUE)
ratings.test = read.csv("/Users/fay_kong/Desktop/SML_SPRING/FILM/test_matchup.csv", head = TRUE)

user.ids = sort(unique(ratings.train$userId))  #count user NO.
movie.ids = sort(unique(ratings.train$movieId)) #count movie NO.

totalmean = mean(ratings.train[,5])


# # # user.bia---------------------------------------------

user.bias = matrix(NA,671,2)
user.bias[,1] = c(1:671)

for(i in 1:671){

  users = ratings.train[which(ratings.train[,1]==i),]

  user.mean = mean(users[,5])

  user.bias[i,2] = user.mean - totalmean
}


# # # movie.bias-------------------------------------------

movie.bias = matrix(NA,8090,3)
movie.bias[,1] = c(1:8090)
movie.bias[,2] = unique(ratings.train$movieId)

for(i in 1:8090){

  movies = ratings.train[which(ratings.train[,2]==movie.bias[i,2]),]

  movie.mean = mean(movies[,5])

  movie.bias[i,3] = movie.mean - totalmean
}


# # # predictions------------------------------------------

ratings.test$prediction = 0

for(i in 1:24720){
  userid = ratings.test[i,1]

  movieid = ratings.test[i,2]

  ubias = user.bias[userid,2]

  movie.row = movie.bias[which(movie.bias[,2]==movieid)]

  if(length(movie.row) == 0){
    mbias = 0
  }else{
    mbias = movie.bias[movie.row,3]
  }

  ratings.test[i,5] = totalmean + ubias +mbias
}

RoundToZeroFive = function(x){
  decimal = x - trunc(x)
  roundx = matrix(NA,length(x),1)
  for(i in 1:length(x)){
    if(decimal[i] < 0.25){
      roundx[i] = trunc(x[i])
    }
    else{
      if(decimal[i] > 0.75){
        roundx[i] = trunc(x[i])+1
      }
      else{
        roundx[i] = trunc(x[i])+0.5
      }
    }
  }
  roundx
}

predictions = RoundToZeroFive(ratings.test[,5])
predictions[which(predictions==0)] = 0.5
predictions[which(predictions > 5)] = 5

write.csv(predictions,file = "/Users/fay_kong/Desktop/SML_SPRING/FILM/Baselinepredictions.csv",row.names = FALSE)