Disclaimer   
Search Trains
 ♦ 
×
DOJ:
Dep:
SunMonTueWedThuFriSat
Class:
2SSLCCEx3AFC2A1A3E
Type:
 

Train Details
Words:
LHB/ICF:
Pantry:
In-Coach Catering/Pantry Car
Loco:
Reversal:
Rake Reversal at Any Stn
Rake:
RSA:
With RSA
Inaug:
 to 
# Halts: to 
Trvl Time: to  (in hrs)
Distance: to  (in kms)
Speed: to  (in km/h)

Departure Details
Include nearby Stations:      ONLY this Station:
Dep Between:    
Dep PF#:
Reversal:
Rake Reversal at Dep Stn

Arrival Details
Include nearby Stations:      ONLY this Station:
Arr Days:
SunMonTueWedThuFriSat
Arr Between:    
Arr PF#:
Reversal:
Rake Reversal at Arr Stn

Search
2-month Availability Calendar
  Go  
Full Site Search
  Search  
 
Sat Oct 1, 2016 19:18:21 ISTHomeTrainsΣChainsAtlasPNRForumGalleryNewsFAQTripsLoginFeedback
Sat Oct 1, 2016 19:18:21 IST
PostPostPost Trn TipPost Trn TipPost Stn TipPost Stn TipAdvanced Search
How the predictions of PNR is calculated?  
1 Answers
Feb 01 2013 (12:28PM)
Site-related

Entry# 1472     
Mayiladuthurai The Undisputed King Of Mainline*^
How the predictions of PNR is calculated?

3808 views
Feb 01 2013 (12:19PM)
Blog Post# 657794-2     
Soumitra Chawathe*^~   Added by: Mayiladuthurai The Undisputed King Of Mainline*^  Feb 01 2013 (12:28PM)
Posting Reply got from Mod sir in pvt, to let others also know this query.
"
PNR Accuracy = No. of Correct Predictions/Total Predictions. Cancelled tickets don't count.
PNR Predictions are separated into 3 slabs:
Current
...
more...
(< 30 days),
Recent (30 - 60 days) and
Old (> 60 days) predictions.
Each of the 3 slabs have their independent PNR Accuracies calculated using the simple formula shown above.
To get the Aggregate accuracy, we average out all 3 of the above numbers, but provide different weightages.
Recent Predictions get twice the weight of Old Predictions.
Current Predictions get twice the weight of Recent Predictions.
.
We have done this because people's PNR Accuracy usually improves with time, as they get more experience, and we do not want their old bad predictions to reduce their overall accuracy %.
.
Example:
Say you have made 200 total predictions out of which 100 were correct.
100 Tot Predictions/60 correct were made before 60 days - OLD
70 Tot Predictions/30 correct within 30-60 days - RECENT
30 Tot Predictions/10 correct within 30 days - CURRENT
Your accuracy % is:
(60*1 + 30*2 + 10*4)/(100*1 + 70*2 + 30*4) = 44%
.
In the example, you have made 50% total correct predictions 100/200. But in the past 30 days, you only got 10/30 correct. Since we give more weight to current predictions, your overall accuracy is only 44%."
.
#pnr
Scroll to Top
Scroll to Bottom


Go to Mobile site