HR ANALYTICS AT SCALENEWORKS - Sec - 2 - Group17

You might also like

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 12

Presented By

Subhranil Mukherjee (FT202087)


Savithru Goutham N (FT202086)
Vrathi J B (FT202099)
Shivangi Gupta (FT202074)
Tanay Banik (FT202093)
Sonam Kala (FT202081)
Sidharth Satyabrat (FT202105)

HR ANALYTICS AT SCALENEWORKS
BEHAVIORAL MODELING TO PREDICT RENEGE
ISSUES

 Supports It companies with its talent acquisition


 Facing difficulties in finding exact match for specified job
 Vicious circle between attrition and recruitment
 Renege is around 15-35 percent after accepting the initial offer
 Repetition of the entire process
 30% revoke rate out of 12000 offer rolled out
 15 man hours/candidate put in by a company. Therefore 54000 man hours wasted by one client alone
 Cost of hiring increases between 10-15 percent
TALC (TALENT ACQUISITION LIFE CYCLE)

Fitment
Sourcing Screening Selection PoFu
and Offer
OBJECTIVES

Key drivers

Set the rules

Predictive model to predict renege


WHAT ARE THE KEY DRIVERS THAT INFLUENCE THE CANDIDATE
JOINING OR NOT JOINING THE COMPANY?

setwd("D:/MBA GLIM/Term3/BA/dataset")
library(caTools)
scalene.df=read.csv("Dataset_of_Scalene _clean.csv")
scalene.df=scalene.df[,-c(1)] #Dropping Candidate Reference, Percent Hike expected and Percent hike offered columns
str(scalene.df)
summary(scalene.df)
#partition
set.seed(123)
sample <- sample.split(scalene.df$Status ,SplitRatio = 0.7)
train.df<- subset(scalene.df,sample == TRUE)
valid.df <- subset(scalene.df, sample == FALSE)
training_logit=glm(Status~.,data=train.df,family=binomial(link="logit"))
training_logit
summary(training_logit)
OUTPUTS
CONCLUSION

Driver Effect
Notice period More the notice period, more is the chance of not joining

Offered Band People offered band E1,E2,E3 has more chance of not joining than E0
band

Candidate source Direct and employee referral have more chance of not joining as
compared to agency-based candidates

LOB Except EAS all others have more chance of not joining

Age Older candidates are more likely to join as compared to younger ones.
PREDICTIVE MODEL TO PREDICT RENEGE

scalene.df=read.csv("Scalene_Updated.csv")
str(scalene.df)
#partition
set.seed(123)
sample <- sample.split(scalene.df$Status_Updated,SplitRatio = 0.7)
train.df<- subset(scalene.df,sample == TRUE)
valid.df <- subset(scalene.df, sample == FALSE)
cv.ct=rpart(Status_Updated~.,data=train.df,method = "class",cp=0.00001,minsplit=10,xval=5)
printcp(cv.ct)
min.cp<-cv.ct$cptable[which.min(cv.ct$cptable[,"xerror"]),"CP"]
min.cp
pruned.ct<-prune(cv.ct,cp=min.cp)
prp(pruned.ct, type = 5)
library(caret)
#generate confusion matrix for training data
train.model=predict(pruned.ct,train.df,type="class")
confusionMatrix(train.model,as.factor(train.df$Status_Updated))
train.model=predict(pruned.ct,train.df,type="class")
valid.model=predict(pruned.ct,valid.df,type="class")
confusionMatrix(valid.model,as.factor(valid.df$Status_Updated))
OUTPUTS

Tree.pdf
RULES

 If Candidate is not relocated, notice period is >= 38 days, Percentage difference in expected and actual salary is
>= -7.1%,Duration to accept offer <26 days & >=4 days, Relevant Experience is < 2 years, then the person will
join
EXAMPLE

Durati Notic
DOJ on to e Offer Pecent hike Percent hike Percent Joinin Candidate Ge Candid Rex Loc A
Candidat Extend accep perio ed expected in offered in difference g relocate nd ate in LO atio g Stat
e Ref ed t offer d band CTC CTC CTC Bonus actual er Source Yrs B n e us

IN
M FR Noi 3 Join
2194323No 1 0E3 114.82 42.86 -33.5No No ale Direct 6 A da 1 ed
THANK YOU

You might also like