machine-learning-challenge-4/baselines.R at master · HackerEarth-Challenges/machine-learning-challenge-4 · 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
path <- "challenge_four/"
setwd(path)


## load data
library(data.table)
train <- fread("train_data.csv")
test <- fread("test_data.csv")

## check data
head(train)
head(test)

## check missing values - no missing values
colSums(is.na(train))
colSums(is.na(test))

## check target class
train[,.N/nrow(train), target]


# Model 1 -----------------------------------------------------------------

## lets predict the majority class
sampsub <- fread("sample_submission.csv")
samplesub[, target := 0]
fwrite(samplesub, "sub0.csv")


# Model 2 + 3 -----------------------------------------------------------------
## Random Forest
## Gradient Boosting

library(xgboost)
library(caTools)

getXGBData <- function(train, test)
{
  split <- sample.split(Y = train$target, SplitRatio = 0.5)

  dtrain <- train[split]
  dvalid <- train[!split]

  dtrain <- xgb.DMatrix(data = as.matrix(dtrain[,-c('connection_id','target'), with=F]), label = dtrain$target)
  dvalid <- xgb.DMatrix(data = as.matrix(dvalid[, -c('connection_id', 'target'), with = F]), label = dvalid$target)
  dtest <- xgb.DMatrix(data = as.matrix(test[,-c('connection_id'), with=F]))

  return (list(train = dtrain, test = dtest, eval = dvalid))

}

getMulAcc <- function(pred, dtrain)
{
  label <- getinfo(dtrain, "label")
  acc <- mean(label == pred)
  return(list(metric = "maccuracy", value = acc))
}


runModels <- function(dtrain, dtest, dvalid, XGB = 0) ## set 1 for XGB: Default run is RF
{

  if(XGB == 1)
  {
    cat('Running Gradient Boosting Model...\n\n')
    # default parameters
    params <- list(objective = 'multi:softmax',
                   num_class = 3)

    watchlist <- list('train' = dtrain, 'valid' = dvalid)
    clf <- xgb.train(params
                     ,dtrain
                     ,1000
                     ,watchlist
                     ,feval = getMulAcc
                     ,print_every_n = 20
                     ,early_stopping_rounds = 30
                     ,maximize = T
    )


    pred <- predict(clf, dtest)

  } else if (XGB == 0) {

    cat('Running Random Forest Model...\n\n')
    params <- list(booster = 'dart'
                   ,objective = 'multi:softmax'
                   ,num_class = 3
                   ,normalize_type = 'tree'
                   ,rate_drop = 0.1)

    watchlist <- list('train' = dtrain, 'valid' = dvalid)
    clf <- xgb.train(params
                     ,dtrain
                     ,1000
                     ,watchlist
                     ,feval = getMulAcc
                     ,print_every_n = 20
                     ,early_stopping_rounds = 30
                     ,maximize = T
    )


    pred <- predict(clf, dtest)

  }

  return(pred)

}

xgbdata <- getXGBData(train, test)

predsRF <- runModels(dtrain = xgbdata$train, dtest = xgbdata$test, dvalid = xgbdata$eval)
predsXGB <- runModels(dtrain = xgbdata$train, dtest = xgbdata$test, dvalid = xgbdata$eval,XGB = 1)


## make submissions
sampsub1 <- fread("sample_submission.csv")
sampsub1[, target := predsRF]
fwrite(sampsub1, "sub1.csv")

sampsub2 <- fread("sample_submission.csv")
sampsub2[, target := predsXGB]
fwrite(sampsub2, "sub2.csv")


## What's next ?
## Since the data set is pretty clean, you can now start with feature engineering, ensembling models etc.