------------------------------------------------------------------------------------------------------------- log: /Users/pabsta/Documents/2-Enseignement/ECON452/tutorial3/OUTPUT/analysis.log log type: text opened on: 3 Feb 2012, 12:49:07 . use "http://www.pabsta.qc.ca/sites/default/files/threeSeries.dta" . . tset t time variable: t, 1 to 251 delta: 1 unit . . /* > Basics of the Box-Jenkins methodology: > 1- Is the process stationary? If no -->Êwe don't know what to do yet. > If yes ->Êproceed to find candidates "p & q" for an ARMA(p,q) representation. > 2- Analysis of PAC, AC graphs to find upper bounds to p & q. Look for particuliar patterns > 3- Proceed to ARIMA estimation and check if we can deduce the process through statistical test > 4- Choose plausible models based on its ability to forecast and other factors. (not seen in class yet). > */ . . /* S1 is the process stationary? */ . tsline s1 if(t<100) // < 100 so we can see something . . //Looks like it oscilates around a stable average -> indication of stationarity. . //Proceed to PAC, AC analysis: . ac s1 . . //Two things to deduce from this graph: . /* First, only the second lag seems to be "spiking out" and is significantly different than zero. > This is an indication that that there is probably an MA(2) component in the process. > Second, there seems to be a pattern of oscillations in the lags (going up and down) > This is an indication that there might be an hidden AR process in there. > */ . pac s1 . /* Pretty much the same story here, but "inversed". There is a spike in the second lag, suggesting > an AR(2) component in the process. There also seem to be an oscillation pattern, which might hide a > n MA process. > > HENCE, we should start our broad estimation with and ARMA(2,2) process. > */ . . //Same thing as both previous graphs, but with an onscreen output. . corrgram s1 -1 0 1 -1 0 1 LAG AC PAC Q Prob>Q [Autocorrelation] [Partial Autocor] ------------------------------------------------------------------------------- 1 -0.0313 -0.0313 .24827 0.6183 | | 2 0.2689 0.2689 18.695 0.0001 |-- |-- 3 -0.0256 -0.0113 18.863 0.0003 | | 4 0.0995 0.0238 21.41 0.0003 | | 5 -0.0492 -0.0357 22.036 0.0005 | | 6 0.0085 -0.0263 22.054 0.0012 | | 7 -0.0824 -0.0639 23.823 0.0012 | | 8 -0.0335 -0.0357 24.116 0.0022 | | 9 -0.0029 0.0418 24.119 0.0041 | | 10 -0.0385 -0.0218 24.509 0.0064 | | 11 0.0569 0.0571 25.366 0.0081 | | 12 0.0154 0.0380 25.429 0.0129 | | 13 -0.0064 -0.0501 25.44 0.0202 | | 14 -0.0209 -0.0439 25.557 0.0294 | | 15 -0.0449 -0.0581 26.101 0.0370 | | 16 -0.0606 -0.0603 27.094 0.0405 | | 17 -0.0890 -0.0824 29.242 0.0324 | | 18 -0.0697 -0.0428 30.568 0.0323 | | 19 -0.0743 -0.0253 32.079 0.0306 | | 20 -0.0012 0.0267 32.08 0.0425 | | 21 0.0040 0.0471 32.084 0.0574 | | 22 0.0727 0.0817 33.552 0.0545 | | 23 0.1166 0.1212 37.336 0.0300 | | 24 0.0194 -0.0206 37.442 0.0395 | | 25 0.0997 0.0415 40.237 0.0276 | | 26 0.0222 0.0157 40.376 0.0358 | | 27 0.0541 0.0014 41.207 0.0393 | | 28 0.0256 0.0377 41.394 0.0494 | | 29 -0.0413 -0.0815 41.882 0.0575 | | 30 0.0907 0.1093 44.247 0.0453 | | 31 -0.0464 -0.0396 44.868 0.0512 | | 32 0.0467 -0.0128 45.499 0.0575 | | 33 -0.0710 -0.0715 46.967 0.0545 | | 34 -0.0028 -0.0748 46.97 0.0685 | | 35 -0.0380 -0.0085 47.394 0.0788 | | 36 -0.0263 -0.0349 47.599 0.0935 | | 37 -0.0625 -0.0431 48.757 0.0935 | | 38 -0.0711 -0.0551 50.264 0.0879 | | 39 0.0091 0.0750 50.289 0.1063 | | 40 -0.0887 -0.0240 52.659 0.0867 | | . . //First estimation of an ARMA(2,2) model: . /* > The command below asks stata "perform a maximum likelihood estimation on the following model: > > s1 = cons + a1 L.s1 + a2 L2.s1 + err + b1 L.err + b2 L2.err > > and give us the estimates" > */ . arima s1, ar(1,2) ma(1,2) (setting optimization to BHHH) Iteration 0: log likelihood = -356.30611 Iteration 1: log likelihood = -354.27742 Iteration 2: log likelihood = -354.06593 Iteration 3: log likelihood = -354.04218 Iteration 4: log likelihood = -354.03837 (switching optimization to BFGS) Iteration 5: log likelihood = -354.03744 Iteration 6: log likelihood = -354.03703 Iteration 7: log likelihood = -354.03703 ARIMA regression Sample: 1 - 251 Number of obs = 251 Wald chi2(4) = 25.40 Log likelihood = -354.037 Prob > chi2 = 0.0000 ------------------------------------------------------------------------------ | OPG s1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- s1 | _cons | 20.20444 .0848579 238.10 0.000 20.03812 20.37076 -------------+---------------------------------------------------------------- ARMA | ar | L1. | -.0756083 .2145822 -0.35 0.725 -.4961817 .344965 L2. | .348233 .2264562 1.54 0.124 -.0956129 .792079 ma | L1. | .0579677 .2207284 0.26 0.793 -.374652 .4905873 L2. | -.0850186 .2457602 -0.35 0.729 -.5666997 .3966624 -------------+---------------------------------------------------------------- /sigma | .9912897 .0440894 22.48 0.000 .9048761 1.077703 ------------------------------------------------------------------------------ . /* > None of the coefficients are independently significant (zero belongs to the confidence interval everywhere) > , > but if we look at the F test of the joint restriction that they are all equal to zero (F test in the upper > right corner), > we see that the p-value is close to zero, which indicates that some variables are somehow significant. > > Lets test if the MA component is jointly insignificant: > */ . test ([ARMA]L.ma = 0) ([ARMA]L2.ma = 0) ( 1) [ARMA]L.ma = 0 ( 2) [ARMA]L2.ma = 0 chi2( 2) = 0.22 Prob > chi2 = 0.8967 . /* > The p-value of this test is 0.89, so we cannot reject the null hypothesis that both coefficients are equal > to zero. > > Before we proceed to estimate the process without an MA component, lets test that the AR component is joint > ly equal to zero. > */ . test ([ARMA]L.ar = 0) ([ARMA]L2.ar = 0) ( 1) [ARMA]L.ar = 0 ( 2) [ARMA]L2.ar = 0 chi2( 2) = 2.85 Prob > chi2 = 0.2401 . /* > p-value of 0.24 (higher than 5%) so we also keep h0. > > The two tests contradict themselves somehow. One suggest that we should drop the AR component while the oth > er says we should > drop the MA component. We should proceed with additional tests before going further. > */ . . test ([ARMA]L.ar = 0) ([ARMA]L.ma = 0) //Test if the process is solely an AR(2), MA(2) ( 1) [ARMA]L.ar = 0 ( 2) [ARMA]L.ma = 0 chi2( 2) = 0.19 Prob > chi2 = 0.9074 . test ([ARMA]L.ar = 0) ([ARMA]L.ma = 0) ([ARMA]L2.ma = 0) //Test if the process is solely an AR(2) ( 1) [ARMA]L.ar = 0 ( 2) [ARMA]L.ma = 0 ( 3) [ARMA]L2.ma = 0 chi2( 3) = 0.39 Prob > chi2 = 0.9419 . test ([ARMA]L.ar = 0) ([ARMA]L.ma = 0) ([ARMA]L2.ar = 0) //Test if the process is solely an MA(2) ( 1) [ARMA]L.ar = 0 ( 2) [ARMA]L.ma = 0 ( 3) [ARMA]L2.ar = 0 chi2( 3) = 3.00 Prob > chi2 = 0.3919 . . /* > These tests suggest that we have might have either an ar(2) process or an ma(2) process. > */ . . //This model is a candidate solution to the process at hand. . arima s1, ar(2) (setting optimization to BHHH) Iteration 0: log likelihood = -354.22133 Iteration 1: log likelihood = -354.21179 Iteration 2: log likelihood = -354.21174 Iteration 3: log likelihood = -354.21174 ARIMA regression Sample: 1 - 251 Number of obs = 251 Wald chi2(1) = 23.18 Log likelihood = -354.2117 Prob > chi2 = 0.0000 ------------------------------------------------------------------------------ | OPG s1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- s1 | _cons | 20.20459 .0855724 236.11 0.000 20.03687 20.3723 -------------+---------------------------------------------------------------- ARMA | ar | L2. | .2725027 .0566023 4.81 0.000 .1615641 .3834412 -------------+---------------------------------------------------------------- /sigma | .9919883 .0431874 22.97 0.000 .9073424 1.076634 ------------------------------------------------------------------------------ . . //Here a new variable res1_1 is built . predict res1_1, r . //The variable contains the residuals (estimated error terms) . //This is a statistical test to see of the residuals are somehow a white noise (we have not seen this test > yet) . wntestq res1_1 Portmanteau test for white noise --------------------------------------- Portmanteau (Q) statistic = 22.0573 Prob > chi2(40) = 0.9905 . /*High p-value: indication that we should keep h0*/ . //This model is also a candidate solution to the process at hand. . arima s1, ma(2) (setting optimization to BHHH) Iteration 0: log likelihood = -355.55437 Iteration 1: log likelihood = -355.2815 Iteration 2: log likelihood = -355.27871 Iteration 3: log likelihood = -355.27866 Iteration 4: log likelihood = -355.27866 ARIMA regression Sample: 1 - 251 Number of obs = 251 Wald chi2(1) = 16.30 Log likelihood = -355.2787 Prob > chi2 = 0.0001 ------------------------------------------------------------------------------ | OPG s1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- s1 | _cons | 20.20281 .078254 258.17 0.000 20.04943 20.35618 -------------+---------------------------------------------------------------- ARMA | ma | L2. | .2451739 .0607324 4.04 0.000 .1261407 .3642071 -------------+---------------------------------------------------------------- /sigma | .9962722 .0429808 23.18 0.000 .9120314 1.080513 ------------------------------------------------------------------------------ . predict res1_2, r . wntestq res1_1 Portmanteau test for white noise --------------------------------------- Portmanteau (Q) statistic = 22.0573 Prob > chi2(40) = 0.9905 . . /* > We still have to pick one of the two models we have found so far. We do not know yet how to do this, but if > a model > is better at forecasting, this should be taken as an indication that the process is somehow "better. > > To do this, we check the variance of the residuals. If the variance is higher, the forecasting error are gr > eater and thus, > the model is less performant. > */ . sum res1_1 res1_2 Variable | Obs Mean Std. Dev. Min Max -------------+-------------------------------------------------------- res1_1 | 251 -.0018225 .9947077 -2.952494 2.925214 res1_2 | 251 -.0011087 .9988701 -2.919931 2.963723 . /* The first model has a slightly smaller variance so perhaps we should keep it.*/ . . . //Second series . tsline s2 if(t<100) . ac s2 . pac s2 . corrgram s2 -1 0 1 -1 0 1 LAG AC PAC Q Prob>Q [Autocorrelation] [Partial Autocor] ------------------------------------------------------------------------------- 1 0.5965 0.5966 90.386 0.0000 |---- |---- 2 0.0060 -0.5457 90.395 0.0000 | ----| 3 -0.3173 0.0012 116.18 0.0000 --| | 4 -0.3172 -0.0522 142.04 0.0000 --| | 5 -0.1747 -0.0770 149.92 0.0000 -| | 6 0.0038 0.0677 149.92 0.0000 | | 7 0.1450 0.0509 155.4 0.0000 |- | 8 0.1087 -0.1538 158.49 0.0000 | -| 9 0.0127 0.0895 158.53 0.0000 | | 10 -0.0520 -0.0379 159.24 0.0000 | | 11 -0.0724 -0.0459 160.63 0.0000 | | 12 -0.0840 -0.0366 162.5 0.0000 | | 13 -0.0859 -0.0532 164.47 0.0000 | | 14 -0.0724 -0.0747 165.88 0.0000 | | 15 -0.0611 -0.0369 166.88 0.0000 | | 16 -0.0076 0.0461 166.9 0.0000 | | 17 0.0626 -0.0069 167.96 0.0000 | | 18 0.0703 -0.0550 169.31 0.0000 | | 19 -0.0004 -0.0354 169.31 0.0000 | | 20 -0.0729 -0.0315 170.77 0.0000 | | 21 -0.0816 -0.0081 172.61 0.0000 | | 22 -0.0679 -0.0656 173.89 0.0000 | | 23 -0.0434 -0.0350 174.41 0.0000 | | 24 0.0180 0.0369 174.5 0.0000 | | 25 0.0362 -0.1094 174.87 0.0000 | | 26 0.0568 0.1266 175.78 0.0000 | |- 27 0.1249 0.1320 180.21 0.0000 | |- 28 0.1830 0.0363 189.74 0.0000 |- | 29 0.1134 -0.0254 193.42 0.0000 | | 30 -0.0148 0.0661 193.48 0.0000 | | 31 -0.0671 0.0381 194.78 0.0000 | | 32 -0.0649 -0.0014 196 0.0000 | | 33 -0.0500 -0.0029 196.73 0.0000 | | 34 -0.0125 0.0070 196.78 0.0000 | | 35 0.0561 0.0484 197.71 0.0000 | | 36 0.0634 -0.0054 198.89 0.0000 | | 37 -0.0085 -0.0472 198.91 0.0000 | | 38 -0.0707 0.0017 200.41 0.0000 | | 39 -0.1070 -0.0762 203.84 0.0000 | | 40 -0.1336 -0.0876 209.21 0.0000 -| | . . /* > The PAC command reveals a clear and bright picture. Both the first lag and the second lag spikes out while > the > remaining lags seems to be non-significant withouth any distinguishable pattern. This is an indication that > the > process at hand is an AR(2) with both lags (first and second) being important. > > On the other hand, the AC command suggest that the AR model dominates since there are clear patterns of osc > illation. > Nonetheless, we will keep up to four lags in our first estimation, simply to encompass the whole possibilit > ies. > > Both have exponential decay, so they also suggest stationarity. > */ . arima s2, ar(1,2) ma(1,2, 3, 4) (setting optimization to BHHH) Iteration 0: log likelihood = -362.35339 Iteration 1: log likelihood = -361.6368 Iteration 2: log likelihood = -361.54328 Iteration 3: log likelihood = -361.40068 Iteration 4: log likelihood = -361.3617 (switching optimization to BFGS) Iteration 5: log likelihood = -361.29259 Iteration 6: log likelihood = -361.17544 Iteration 7: log likelihood = -361.09042 Iteration 8: log likelihood = -361.07164 Iteration 9: log likelihood = -361.06925 Iteration 10: log likelihood = -361.06908 Iteration 11: log likelihood = -361.06902 Iteration 12: log likelihood = -361.06901 Iteration 13: log likelihood = -361.06901 ARIMA regression Sample: 1 - 251 Number of obs = 251 Wald chi2(6) = 362.95 Log likelihood = -361.069 Prob > chi2 = 0.0000 ------------------------------------------------------------------------------ | OPG s2 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- s2 | _cons | -10.06652 .1049567 -95.91 0.000 -10.27223 -9.86081 -------------+---------------------------------------------------------------- ARMA | ar | L1. | 1.158292 .2169143 5.34 0.000 .7331478 1.583436 L2. | -.6251838 .1733191 -3.61 0.000 -.9648831 -.2854846 ma | L1. | -.2391491 .234635 -1.02 0.308 -.6990253 .2207272 L2. | -.1339535 .1403526 -0.95 0.340 -.4090396 .1411325 L3. | -.0567033 .1358303 -0.42 0.676 -.3229258 .2095192 L4. | .1651303 .1310021 1.26 0.207 -.0916292 .4218897 -------------+---------------------------------------------------------------- /sigma | 1.017297 .0492341 20.66 0.000 .9208002 1.113794 ------------------------------------------------------------------------------ . test ([ARMA]L3.ma = 0) ([ARMA]L2.ma = 0) ( 1) [ARMA]L3.ma = 0 ( 2) [ARMA]L2.ma = 0 chi2( 2) = 0.95 Prob > chi2 = 0.6216 . arima s2, ar(1,2) ma(1, 4) (setting optimization to BHHH) Iteration 0: log likelihood = -362.40997 Iteration 1: log likelihood = -361.76211 Iteration 2: log likelihood = -361.6814 Iteration 3: log likelihood = -361.67435 Iteration 4: log likelihood = -361.6737 (switching optimization to BFGS) Iteration 5: log likelihood = -361.67365 Iteration 6: log likelihood = -361.67364 ARIMA regression Sample: 1 - 251 Number of obs = 251 Wald chi2(4) = 389.38 Log likelihood = -361.6736 Prob > chi2 = 0.0000 ------------------------------------------------------------------------------ | OPG s2 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- s2 | _cons | -10.06621 .1137148 -88.52 0.000 -10.28909 -9.843335 -------------+---------------------------------------------------------------- ARMA | ar | L1. | 1.103689 .1393275 7.92 0.000 .8306123 1.376766 L2. | -.6713611 .0964976 -6.96 0.000 -.8604929 -.4822292 ma | L1. | -.1814701 .1612964 -1.13 0.261 -.4976054 .1346651 L4. | .161433 .1054755 1.53 0.126 -.0452951 .3681611 -------------+---------------------------------------------------------------- /sigma | 1.019786 .0493092 20.68 0.000 .9231413 1.11643 ------------------------------------------------------------------------------ . arima s2, ar(1,2) ma(1) (setting optimization to BHHH) Iteration 0: log likelihood = -362.69039 Iteration 1: log likelihood = -362.66453 Iteration 2: log likelihood = -362.66356 Iteration 3: log likelihood = -362.66351 Iteration 4: log likelihood = -362.66351 ARIMA regression Sample: 1 - 251 Number of obs = 251 Wald chi2(3) = 279.22 Log likelihood = -362.6635 Prob > chi2 = 0.0000 ------------------------------------------------------------------------------ | OPG s2 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- s2 | _cons | -10.06682 .1069846 -94.10 0.000 -10.2765 -9.857132 -------------+---------------------------------------------------------------- ARMA | ar | L1. | .9168512 .0970015 9.45 0.000 .7267317 1.106971 L2. | -.5413305 .0722583 -7.49 0.000 -.6829542 -.3997069 ma | L1. | .0027421 .1150565 0.02 0.981 -.2227645 .2282488 -------------+---------------------------------------------------------------- /sigma | 1.023979 .0492696 20.78 0.000 .9274126 1.120546 ------------------------------------------------------------------------------ . arima s2, ar(1,2) //Most compact representation. (setting optimization to BHHH) Iteration 0: log likelihood = -362.67043 Iteration 1: log likelihood = -362.66389 Iteration 2: log likelihood = -362.66373 Iteration 3: log likelihood = -362.66373 ARIMA regression Sample: 1 - 251 Number of obs = 251 Wald chi2(2) = 279.79 Log likelihood = -362.6637 Prob > chi2 = 0.0000 ------------------------------------------------------------------------------ | OPG s2 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- s2 | _cons | -10.06689 .106284 -94.72 0.000 -10.2752 -9.858578 -------------+---------------------------------------------------------------- ARMA | ar | L1. | .9187994 .0550185 16.70 0.000 .8109651 1.026634 L2. | -.5425169 .0558461 -9.71 0.000 -.6519732 -.4330607 -------------+---------------------------------------------------------------- /sigma | 1.023938 .0489107 20.93 0.000 .9280751 1.119801 ------------------------------------------------------------------------------ . . predict res2, r . wntestq res2 Portmanteau test for white noise --------------------------------------- Portmanteau (Q) statistic = 43.2427 Prob > chi2(40) = 0.3346 . . //Third series . tsline s3 if(t <100) . /* > Woah! No way this is stationarry! The process is ever increasing with time! > */ . . ac s3 . //No exponential decay. . . pac s3 . //Huge spike in the first lag, suggesting that we have an AR(1) with the coefficient being equal to one. . //This is obviously not a stationarry process. . corrgram s3 -1 0 1 -1 0 1 LAG AC PAC Q Prob>Q [Autocorrelation] [Partial Autocor] ------------------------------------------------------------------------------- 1 0.9871 0.9979 247.48 0.0000 |------- |------- 2 0.9742 -0.0420 489.51 0.0000 |------- | 3 0.9612 -0.0134 726.09 0.0000 |------- | 4 0.9482 -0.0649 957.22 0.0000 |------- | 5 0.9351 0.0200 1183 0.0000 |------- | 6 0.9220 -0.0233 1403.3 0.0000 |------- | 7 0.9089 0.0197 1618.3 0.0000 |------- | 8 0.8958 0.1050 1828 0.0000 |------- | 9 0.8828 0.0808 2032.5 0.0000 |------- | 10 0.8700 0.0069 2231.9 0.0000 |------ | 11 0.8574 -0.0252 2426.4 0.0000 |------ | 12 0.8447 0.0159 2616 0.0000 |------ | 13 0.8318 -0.0074 2800.6 0.0000 |------ | 14 0.8189 -0.0612 2980.3 0.0000 |------ | 15 0.8063 -0.0221 3155.2 0.0000 |------ | 16 0.7937 0.0333 3325.5 0.0000 |------ | 17 0.7810 -0.0698 3491 0.0000 |------ | 18 0.7684 -0.0695 3651.9 0.0000 |------ | 19 0.7560 -0.0319 3808.3 0.0000 |------ | 20 0.7438 -0.0008 3960.4 0.0000 |----- | 21 0.7316 -0.0640 4108.2 0.0000 |----- | 22 0.7194 -0.0805 4251.7 0.0000 |----- | 23 0.7072 -0.0592 4391 0.0000 |----- | 24 0.6950 -0.1093 4526.1 0.0000 |----- | 25 0.6829 0.0288 4657.2 0.0000 |----- | 26 0.6708 -0.0674 4784.2 0.0000 |----- | 27 0.6589 0.0221 4907.3 0.0000 |----- | 28 0.6471 -0.0609 5026.5 0.0000 |----- | 29 0.6354 -0.0598 5142 0.0000 |----- | 30 0.6237 0.0342 5253.8 0.0000 |---- | 31 0.6118 -0.0377 5361.8 0.0000 |---- | 32 0.6002 0.1087 5466.3 0.0000 |---- | 33 0.5886 0.0272 5567.2 0.0000 |---- | 34 0.5771 0.1247 5664.7 0.0000 |---- | 35 0.5657 -0.0063 5758.7 0.0000 |---- | 36 0.5544 0.0526 5849.5 0.0000 |---- | 37 0.5432 -0.0441 5937.1 0.0000 |---- | 38 0.5322 0.0148 6021.5 0.0000 |---- | 39 0.5211 0.0119 6102.9 0.0000 |---- | 40 0.5101 0.0609 6181.2 0.0000 |---- | . . //We try to differenciate the series to see if its stationary: . gen diff_s3 = D.s3 (1 missing value generated) . tsline diff_s3 if(t<100) . . ac diff_s3 . pac diff_s3 . //Seems stationary. It also seem we have no lag explaining the difference. . /* > This means that the first difference of the series is somehow a white noise. > */ . . wntestq diff_s3 Portmanteau test for white noise --------------------------------------- Portmanteau (Q) statistic = 38.5701 Prob > chi2(40) = 0.5347 . //This is also suggested by this test. . //The statistics are : . sum diff_s3 Variable | Obs Mean Std. Dev. Min Max -------------+-------------------------------------------------------- diff_s3 | 250 1.094539 .9899983 -2.694153 3.912323 . /* This suggest that the model is given by > > y = 1.09 + L.y + u where u ~N(0, 1) > */ . log close log: /Users/pabsta/Documents/2-Enseignement/ECON452/tutorial3/OUTPUT/analysis.log log type: text closed on: 3 Feb 2012, 12:49:36 -------------------------------------------------------------------------------------------------------------