Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
prediction test
Message posted by ansuman (via 210.214.202.179) on January 6, 2002 at 11:22 PM (ET)
*****please maximize the notpad to get proper view of the table. *****
problem:
i have given a table.in this table by the use of six inputs we get the output.but it takes two hours to get the output. can we by means of statistics ( suppose i have 200 previous data )predict the output after inputing the inputs without waiting for the processing..
the last line of the table if we enter the last line inputs can we get the output result.
it will be very kind if anybody helps me in this .i am requesting all the statistics gurus to please help me out.
if anybody can help theni will be very greatfull towards him.
output input1 input2 input3 input4 input6
1470.05 0.6200000048 85.00015258789 35.90999984741 0.0671134442 4.2055759430
1472.87 0.6209463477 82.26740264893 35.90999984741 0.0680355281 4.1389207840
1472.87 0.6455397606 80.00122070313 35.90999984741 0.0745448917 4.1661930084
1472.87 0.6719470024 80.00122070313 35.90999984741 0.0817084089 4.2005834579
1471.95 0.6984622478 80.00122070313 35.90999984741 0.0887810513 4.2350502014
1461.94 0.7556277514 84.13645172119 35.90999984741 0.0705082342 4.2911343575
1461.94 0.8060192466 85.00015258789 35.90999984741 0.0454391874 4.3530020714
1461.94 0.7991335392 85.00015258789 35.90999984741 0.0446027219 4.3941268921
1461.94 0.7909683585 82.26740264893 35.90999984741 0.0443043262 4.4347906113
1479.39 0.7821747065 80.00122070313 35.90999984741 0.0441916287 4.4738707542
1491.82 0.7358129025 80.00122070313 35.90999984741 0.0551009960 4.4186859131
1491.82 0.6600562334 80.00122070313 35.90999984741 0.0665742457 4.2818083763
1491.82 0.6125071645 80.00122070313 35.90999984741 0.0622085854 4.1826057434
1491.82 0.6159377098 80.00122070313 35.90999984741 0.0749358609 4.2272024155
1491.82 0.6447387934 80.00122070313 35.90999984741 0.0835554600 4.2674965858
1491.82 0.9209833145 80.00122070313 35.90999984741 0.0495544747 4.2526359558
1491.82 0.8229721785 80.00122070313 35.90999984741 0.0645173490 4.2450385094
1491.82 0.6812810898 80.00122070313 35.90999984741 0.0860791355 4.2360267639
1491.82 0.6628953815 80.00122070313 35.90999984741 0.0968690440 4.1794371605
1487.54 0.6815100908 80.40113830566 35.90999984741 0.1031554788 4.1248049736
1441.09 0.7456413507 85.00015258789 35.90999984741 0.0897413120 4.2408246994
1444.31 0.7892169952 85.00015258789 35.90999984741 0.0783821270 4.3000001907
1444.31 0.8019412756 85.28897857666 35.90999984741 0.0763107389 4.3000001907
1447.85 0.8003721237 89.99908447266 35.90999984741 0.0790571272 4.2643857002
1486.5 0.7744144201 89.99908447266 35.90999984741 0.0902387872 4.1827383041
1486.5 0.7762581110 89.99908447266 35.90999984741 0.0952444226 4.2063350677
************************************************************************************
*1486.5* 0.7989953756 89.99908447266 35.90999984741 0.0941950232 4.24306488040
*************************************************************************************
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: prediction test
Message posted by JG (via 172.173.212.7) on January 7, 2002 at 6:58 AM (ET)
You can use time series methods but it will not be the same, because you will be approximating a complex model with a much simpler one.
Re: prediction test
Message posted by Rich (via 62.252.32.6) on January 17, 2002 at 7:43 AM (ET)
You could always try neural networks, provided you have sufficient data to train one.
Re: prediction test
Message posted by Phil (via 66.32.16.14) on January 18, 2002 at 12:28 AM (ET)
Probably the simplest way is to build a multiple regression model with a good "fit", that is one whose residuals are normally distributed with constant variance over the range of the dependent variable. Hopefully you would not have to do any data transformations to get a good fit.Look in any basic statistics text under "Multiple Regression".
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.
