For discussions related to modeling, machine learning and deep learning. Related packages include caret, modelr, yardstick, rsample, parsnip, tensorflow, keras, cloudml, and tfestimators.
Hi All,
I am trying to build anomaly detection code. But my code is taking a huge time. Please let me know how can i speed up the process .
Thanks in advance for your help
### Please find below code
##model creation
dataFile <- read.csv("filepth\xx.txt",header = FALSE, sep = ";")
dataFile$TimeTag <-as.POSIXct(paste(dataFile$V1,dataFile$V2))
for creating model by training data on 25000 rows
dataFile <- dataFile[0:25000,]
75% of the sample size
smp_size <- floor(0.80 * nrow(dataFile))
set the seed to make your partition reproducible
set.seed(123)
train_ind <- sample(seq_len(nrow(dataFile)), size = smp_size)
#df <- df[ -c(1, 3:6, 12) ]
train <- dataFile[train_ind, ]
test <- dataFile[-train_ind, ]
##training data consists of all 14 columns
train1 <- train[-c(1,2,15:16)]
##model creation which will be loaded while prediction
res.fpc1 <- fpc::dbscan(train1, eps = 0.1, MinPts = 5,scale =FALSE,method = c("hybrid"))
prediction
library(shiny)
library(shinydashboard)
library(DT)
library(data.table)
library(shinyalert)
library(shinycssloaders)
library(stringr)
library(dygraphs)
library(datasets)
library(xts)
#library(mlbench)
library(caret)
library(dbscan)
library(fpc)
library(lubridate)
t1<-Sys.time()
dataFile <- read.csv("filepathxxxx",header = FALSE, sep = ";")
##training data
dataFile_train <- dataFile[0:25000,]
dataFile$TimeTag <-as.POSIXct(paste(dataFile$V1,dataFile$V2))
newdata <- dataFile
dataFile1 <- dataFile[-c(1,2,15)]
75% of the sample size
smp_size <- floor(0.80 * nrow(dataFile))
set the seed to make your partition reproducible
set.seed(123)
train_ind <- sample(seq_len(nrow(dataFile)), size = smp_size)
#df <- df[ -c(1, 3:6, 12) ]
##predicting for 5th and 7 th column
dataFile_train <- dataFile_train[-c(1,2,3,4,6,8:16)]
newdata1 <- newdata[-c(1,2,3,4,6,8:16)]
##predicting for 5th columnonly
#newdata1 <- newdata[-c(1:4,6:16)]
#dataFile_train <- dataFile_train[-c(1:4,6:16)]
#loading model
model2 <- readRDS("\filepath\xxx.rda")
##prediction
y_pred<- predict.dbscan(res.fpc1,dataFile_train, newdata = newdata1)
newdata1["Predictions"] = y_pred
newdata1["time"] = newdata[,"TimeTag"]
anomalyDataframe <- subset(newdata1, Predictions == 0)