Hello,   I am building a model to find those customers which are likely to be dormant in future. We want our model to have least misclassification rate for better campaigning activity.   Through Decision Tree my Misclassification Rate is 0.29. To reduce MR I have used Gradient Boosting, as I have gone through various papers GB reduces MR with multiple iteration, but even GB is showing MR 0.29.   I am not able to understand.   1) How is Gradient Boosting different from Decision Tree? 2) When to use Gradient Boosting, Decision Tree & Logistic Regression? 3) How to reduce misclassification rate?   Thank You, Vishal ... View more

Dear All, Greetings for the Day..!! Well I have trade level data for 50 days. For each day, there are 5 columns (namely Day, Flag, Price, Qty, Type) . Since my data is for 50 days, which means there are total 250 columns. (There are 60,000 plus rows for each day.) Data: 8000 Call 8100 Call 8200 Call Day Flag Price Qty Type Day Flag Price Qty Type Day Flag Price Qty Type May 4, 2015 TRADE 320 4 May 5, 2015 TRADE 368.05 1 May 6, 2015 TRADE 355 5 May 4, 2015 TRADE 320.1 2 May 5, 2015 TRADE 368 3 May 6, 2015 TRADE 354.25 2 May 4, 2015 TRADE 316.85 1 May 6, 2015 TRADE 368 4 May 6, 2015 TRADE 354.3 2 May 4, 2015 TRADE 317.4 1 May 6, 2015 TRADE 368 2 May 6, 2015 TRADE 354.2 1 May 4, 2015 TRADE 311.05 1 May 7, 2015 TRADE 361.65 2 May 6, 2015 TRADE 354.4 1 May 4, 2015 TRADE 320.3 18 May 7, 2015 TRADE 362.8 7 May 6, 2015 TRADE 350.05 1 May 4, 2015 TRADE 318 18 May 8, 2015 TRADE 361.85 3 May 6, 2015 TRADE 347.05 2 May 4, 2015 TRADE 318 1 May 8, 2015 TRADE 362.8 1 May 6, 2015 TRADE 345.85 7 May 4, 2015 TRADE 318 2 May 9, 2015 TRADE 362.15 8 May 6, 2015 TRADE 342 3 May 4, 2015 TRADE 318 2 May 9, 2015 TRADE 366.05 1 May 6, 2015 TRADE 341 131 May 4, 2015 TRADE 318.05 2 May 10, 2015 TRADE 368 1 May 6, 2015 TRADE 343 12 May 4, 2015 TRADE 318 7 May 10, 2015 TRADE 368 3 May 6, 2015 TRADE 343 10 May 4, 2015 TRADE 320 40 May 11, 2015 TRADE 367 8 May 6, 2015 TRADE 345.1 1 Desired Output: Day Flag Price Qty Type May 4, 2015 8000 320 4 Call May 4, 2015 8000 320.1 2 Call May 4, 2015 8000 316.85 1 Call May 4, 2015 8000 317.4 1 Call May 4, 2015 8000 311.05 1 Call May 4, 2015 8000 320.3 18 Call May 4, 2015 8000 318 18 Call May 4, 2015 8000 318 1 Call May 4, 2015 8000 318 2 Call May 4, 2015 8000 318 2 Call May 4, 2015 8000 318.05 2 Call May 4, 2015 8000 318 7 Call May 4, 2015 8000 320 40 Call May 5, 2015 8100 368.05 1 Call May 5, 2015 8100 368 3 Call May 5, 2015 8100 368 4 Call May 5, 2015 8100 368 2 Call May 5, 2015 8100 361.65 2 Call May 5, 2015 8100 362.8 7 Call May 5, 2015 8100 361.85 3 Call May 5, 2015 8100 362.8 1 Call May 5, 2015 8100 362.15 8 Call May 5, 2015 8100 366.05 1 Call May 5, 2015 8100 368 1 Call May 5, 2015 8100 368 3 Call May 5, 2015 8100 367 8 Call May 6, 2015 8200 355 5 Call May 6, 2015 8200 354.25 2 Call May 6, 2015 8200 354.3 2 Call May 6, 2015 8200 354.2 1 Call May 6, 2015 8200 354.4 1 Call May 6, 2015 8200 350.05 1 Call May 6, 2015 8200 347.05 2 Call May 6, 2015 8200 345.85 7 Call May 6, 2015 8200 342 3 Call May 6, 2015 8200 341 131 Call May 6, 2015 8200 343 12 Call May 6, 2015 8200 343 10 Call May 6, 2015 8200 345.1 1 Call Stpes required to get desired output: 1. Column Flag should consider first 4 letters of Header present in first row (e.g.8000 Call Header will show flag as 8000) 2. Column Type should consider 6-9 letters of Header (e.g. 8000 Call will show Type as Call) 3. Now,Append each days data one under other. Request to kindly help me to get me desired output. Thanks and Regards, Vishal Kapasi ... View more

Dear All, Greetings for the day... Well, I want to pair my Sell transactions with Buy transactions considering LIFO logic. This means, whenever I sell shares I want to know its corresponding BUY, who were last bought prior to my Sell. For example, On 1 st of July, I bought 100, and 25 shares, and sold all 125 shares, So prior to selling 125 shares, my last bought 125 shares were also on 1st of July. So matching price will consider its weighted buy price in desired output. Please find my Input Data in below table. Trade Date Flag Shares Price 1-Jul-15 Buy 100 505 1-Jul-15 Buy 25 520 1-Jul-15 Sell -125 530 2-Jul-15 Buy 50 480 6-Jul-15 Buy 25 492 8-Jul-15 Sell -25 475 10-Jul-15 Buy 50 503 15-Jul-15 Sell -50 515 21-Jul-15 Sell -50 540 Please find below table for Desired output. Here, in case of 8July Sell, I had bought prior to 8July is  50 shares on 2nd and 25 shares on 6July. Since my last 25 shares bought is for 6July, it will be matched considering LIFO logic and its price will be reflected in my matching price output. Trade Date Flag Shares Price LIFO based Matching Price 1-Jul-15 Buy 100 505 1-Jul-15 Buy 25 520 1-Jul-15 Sell -125 530 (25*520+100*505)/125 2-Jul-15 Buy 50 480 6-Jul-15 Buy 25 492 8-Jul-15 Sell -25 475 25*492/25 10-Jul-15 Buy 50 503 15-Jul-15 Sell -50 515 50*503/50 21-Jul-15 Sell -50 540 50*480/50 Requesting for sharing appropriate code considering this LIFO logic where I can get my desired output. Thanks & Regards, Vishal ... View more