tag:blogger.com,1999:blog-7631116270195175228.post6696012031273652723..comments2024-03-14T04:06:37.165-07:00Comments on Text & Data Mining by practical means: Support Vector Regression (SVR): predict earthquakes through sunspotsCristian Mesianohttp://www.blogger.com/profile/04880057603671195464noreply@blogger.comBlogger8125tag:blogger.com,1999:blog-7631116270195175228.post-5207511500827585612023-07-30T16:04:19.847-07:002023-07-30T16:04:19.847-07:00gümüşhane
bilecik
erzincan
nevşehir
niğde
V2İ<a href="https://gumushane.escortdocs.com/" title="gümüşhane" rel="nofollow">gümüşhane</a><br /><a href="https://bilecik.escortdocs.com/" title="bilecik" rel="nofollow">bilecik</a><br /><a href="https://erzincan.escortdocs.com/" title="erzincan" rel="nofollow">erzincan</a><br /><a href="https://nevsehir.escortdocs.com/" title="nevşehir" rel="nofollow">nevşehir</a><br /><a href="https://nigde.escortdocs.com/" title="niğde" rel="nofollow">niğde</a><br />V2İalimnoreply@blogger.comtag:blogger.com,1999:blog-7631116270195175228.post-18370415216993605832012-03-28T12:37:25.513-07:002012-03-28T12:37:25.513-07:00Sometime I love be a bit provocative :) I think be...Sometime I love be a bit provocative :) I think be provocative is in the nature of data miner (even if the first skill must be the humbleness).<br />BTW, in the post I sowed some clues.<br />For instance have a look at the earthquake series in the post's appendix: the graph itself it's really weird, indeed, in the caption I clearly highlighted that the increasing is really suspicious. <br />-> in the next post I would take this freakpost as an example of what a dataminer should avoid to do!<br />just as anticipation: the overall error in the analysis doesn't lay in the technique (SVR has been correctly implemented, trained and tested) but in the way I used (fortunately intentionally) and in the dataset used.<br />So the SVR model overfits the variables (indeed the test explained well the training set but i didn't do tests over a test set never seen before by the model), and dataset itself is not representative of the domain.<br />...Let's discuss on that in the next post.<br /><br />cheers.Cristian Mesianohttps://www.blogger.com/profile/04880057603671195464noreply@blogger.comtag:blogger.com,1999:blog-7631116270195175228.post-61708684997767996602012-03-28T08:36:50.731-07:002012-03-28T08:36:50.731-07:00Come on guys, is it so difficult to just laugh som...Come on guys, is it so difficult to just laugh sometimes?<br />This blog entry is both hilarious and well explained.<br /><br />How many times have you seen the so called medias reporting dubious correlations along the lines of "people who drink more milk during childhood have more chance of becoming rich when adults"?<br /><br />The choice of the data sets is probably intented to make fun of that.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7631116270195175228.post-7322088892456600262012-03-10T00:12:55.912-08:002012-03-10T00:12:55.912-08:00Hi understand that such topic can cause reactions ...Hi understand that such topic can cause reactions by geologists and expert of quake events. I'm totally open to critics, but I would like once again remark that:<br />--> This post is just a tutorial to explain a regression technique.<br />--> The input data are real but even if the regressor function is able to explain very well (especially with #sunspots input >50) this doesn't mean that it is able to forecast quakes!!<br />--> In the post there's mention about "warning" on scientific validity of this example (I think that I have been transparent and correct).<br /><br />With regard to "French comedies": are your concerns (and sarcasm) related to the SVR technique or are related to example used to explain it?<br /><br />@last anonymous: actually I don't see that the former comment was cryptic: the message has been clearly delivered! :)Cristian Mesianohttps://www.blogger.com/profile/04880057603671195464noreply@blogger.comtag:blogger.com,1999:blog-7631116270195175228.post-46200088922627845242012-03-09T13:27:42.548-08:002012-03-09T13:27:42.548-08:00mhmm. I didn't get the meaning of what you sai...mhmm. I didn't get the meaning of what you said. <br />Could you please be less cryptic?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7631116270195175228.post-29051487195646766822012-03-09T13:09:39.776-08:002012-03-09T13:09:39.776-08:00I don't think it's a joke, at all!
That ...I don't think it's a joke, at all! <br /><br />That stuff is serious! Something like that should have made it into one of the most exhilarating french comedies of all time!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7631116270195175228.post-61346563616315217262012-03-09T12:27:18.503-08:002012-03-09T12:27:18.503-08:00Hi,
As I mentioned in the post, it is just an expe...Hi,<br />As I mentioned in the post, it is just an experiment based on real data.<br />Is this model really able to forecast earthquakes? I don't know... it's unlikely that just a variable can describe a complex event as quakes but looking the results I think that a sort of correlation between the two phenomena is not so far from the reality.<br />...In the next post, I gonna use the regression function to a recent dataset (1989 to 2012) and we will check how reliable it is!<br />Stay tuned!Cristian Mesianohttps://www.blogger.com/profile/04880057603671195464noreply@blogger.comtag:blogger.com,1999:blog-7631116270195175228.post-27754569902934474602012-03-09T12:14:31.767-08:002012-03-09T12:14:31.767-08:00Is this a joke?Is this a joke?Anonymousnoreply@blogger.com