• zmq
• mongodb

• numpy
• scipy
• pyzmq
• pymongo

• prediction

  • pnr : md5 of pnr number
  • boarding_time : time of boarding in ISO date format
  • class : Class of Travel
  • from : from station code
  • to : To station code
  • train : train number
  • waiting_list : Waiting list at the time of prediction
  • hours_before : number of hours before the departure of train from the from station at the time of prediction
  • prediction : number, positive value means ticket will be confirmed and negative value means the opposite, zero is undecided
  • quota : Quota in Indian railways
  • status" : Whether the prediction was success or not, possible values "success" and "failure"

• pnr change

  • date : date in string format, example "03-11-2013"
  • x_val : array of time of checking(number of hours before departure)[ 0, 3, 23, 47, 71, 95, 119, 145 ]
  • y_val : array of waiting list at time of checking [ 43, 51, 59, 83, 99, 111, 122, 127 ]
  • pnr : md5 of pnr number
  • class : Class of Travel
  • from : from station code
  • to : To station code
  • train : train number
  • quota : Quota in Indian railways

Download the .bson files in the db folder and add it to the database.

mongorestore -h localhost:port -d pnr -c prediction path/to/prediction.bson
mongorestore -h localhost:port -d pnr -c pnr_change path/to/pnr_change.bson

virtualenv ~/path/to/virtualenv/folder
source ~/path/to/virtualenv/folder/bin/activate
pip install numpy scipy pyzmq pymongo

make changes to the function, run the prediction accuracy job and track whether it has improved accuracy or not

Success in 1709, Failure in 490 Original Success in 1613, Original Failure in 587