@saoulidisg @LukeDashjr That may not happen just as well. After waves of Big Data work made in last decade, researchers found that there is such a thing as too much data. They found the models to produce spurious correlations and lead decision making astr
@these_last_days @ettingermentum Yep, not even 'can' be, a spurious correlation will emerge in all large enough data sets as the probability a series of related values reiterates is never zero. Effect size is everything for big data. https://t.co/uxyz8mBX
@ShaiCarmi @StevePittelli @doctorveera @DPosthu @euffelmann What he said is true. https://t.co/b3b7M21vW9
@renatrigiorese Better link https://t.co/a5AGQ4sPt4.
Oh my lord this is incredible
RT @Race__Realist: @ent3c @Steve_Sailer We also have this: https://t.co/58oyoZ3clf.
@ent3c @Steve_Sailer We also have this: https://t.co/58oyoZ3clf.
@ishirubi There is another path: @nntaleb came to this from a probability point of view. You can have a computability point of view of the same issue. There is for instance this paper on big data by Calude and Longo https://t.co/5z9toR6NY4
My nomination for The Snuffed Candle Award goes to this paper: https://t.co/NulGV8PI8X
The Deluge of Spurious Correlations in Big Data via @teppofelin https://t.co/z82cauuY8r
The dangers of not using a balanced approach when leveraging big data. Opportunities must be weighted against the tradeoffs of unstructured black boxes.
"Too much information tends to behave like very little information"
RT @VanilaSingh: Very true !
Replace "marriage rate in Kentucky" with "inflammation" and "people who drowned..." with "depression".
Very true !
RT @barrett_larson: @EMARIANOMD @mcisaac_d A good reminder that correlation doesn't necessarily imply causation. While Big Data holds a lot…
RT @barrett_larson: @EMARIANOMD @mcisaac_d A good reminder that correlation doesn't necessarily imply causation. While Big Data holds a lot…
@EMARIANOMD @mcisaac_d A good reminder that correlation doesn't necessarily imply causation. While Big Data holds a lot of promise, it can also lead to wrong statistical inferences. We must watch out for spurious correlations. Need RCTs. @TylerVigen https
@rasmansa @sir_deenicus @RCownie @kareem_carr @necoleman What truth means in a pure noise scenario ? This is what is guaranteed with random large enough datasets https://t.co/itqkoCofdu
PapersOfTheDay Lulz^2 (reminder of spurious correlations offered by various anti-vaxxers and anti-GMO folks) "Glyphosate, neurological diseases – and the scientific method" https://t.co/QwGv6NOqJd "The deluge of spurious correlations in big data" https:
RT @Descartes_Ghost: @ctricot Sur le même sujet, d'un point de vue plus scientifique/technique je recommande la lecture de l'article de Cal…
@JoeSmit84460720 GWAS is awash with spurious correlations, strange assumptions and a naive perspective that with enough data and computational power that informatics can simply skip over over actual biology to build a predictive science. They are wrong. h
@AhsanDeliri Maybe this paper of Calude and Longo will interest you (its a theoretical computer science approach to this point) : https://t.co/SYRPCwX26R
RT @Descartes_Ghost: @ctricot Sur le même sujet, d'un point de vue plus scientifique/technique je recommande la lecture de l'article de Cal…
RT @Descartes_Ghost: @ctricot Sur le même sujet, d'un point de vue plus scientifique/technique je recommande la lecture de l'article de Cal…
@ctricot Sur le même sujet, d'un point de vue plus scientifique/technique je recommande la lecture de l'article de Calude et Longo sur les corrélations et leurs mésusages : https://t.co/SYRPCwX26R
RT @pp0196: @DanGraur In large enough datasets, the pony or horse that will be found is with high probability just shit. https://t.co/zx3Vl…
@DanGraur In large enough datasets, the pony or horse that will be found is with high probability just shit. https://t.co/zx3VlZZMMn
@PrismOfReality @smith_valence @GeorgeHemingto1 @sillyolyou @itsbirdemic @r_a_mckinney @AThrowAwayAcco5 @torinmccabe @rasmansa @NoahCarl90 @EPoe187 https://t.co/zx3VlZZMMn Yeah, including predicting noise.
The suggestion here is that big data sets NECESSARILY yield erroneous correlations (eg: the stock market rises when rice prices fall in Bangkok). A potentially significant finding for predictive analytics. https://t.co/YbvJWJicgh
@zakkohane @harvard_data But without a priori, redigered hypothesis, #DataScience allows for far too much statistical fishing (often by unaware researchers) and then we end up with studies like this: https://t.co/wCYPCnFklv or https://t.co/8c3eJ3Aqyu or ht
RT @CoyneoftheRealm: The Deluge of Spurious Correlations in Big Data: Too much information tends to behave like very little information. Th…
The Deluge of Spurious Correlations in Big Data: Too much information tends to behave like very little information. The scientific method can be enriched by computer mining in immense databases, but not replaced by it. https://t.co/6YrpPaDrsi
@gcochran99 @itsbirdemic @samizstat @joftius @andmisterbill @amiguello1 @negatingspirit @dbweissman @acid_tv @HobbesianM @CrecusLohe @YeyoZa @cgoldberg618 Cases where misleading results will be replicated: https://t.co/6mb8x2Prbg And as the features spac
@BBCPallab Or to put it in the more formal way: https://t.co/4US8PkL83s
RT @trentyarwood: This is an extremely Trent article. https://t.co/GatgWhH8Mq #bigdata (ping @jpwarren)
RT @trentyarwood: This is an extremely Trent article. https://t.co/GatgWhH8Mq #bigdata (ping @jpwarren)
This is an extremely Trent article. https://t.co/GatgWhH8Mq #bigdata (ping @jpwarren)
@f2harrell @LauraBBalzer @EpiEllie https://t.co/RbUrhPk4kF If the features space is not that big.
Today's reading: Longo & Calude, "The Deluge of Spurious Correlations in Big Data" - https://t.co/PiUjraw6RM #bigdata
RT @Descartes_Ghost: @nntaleb Two references on this subject : "The deluge of spurious correlations in big data" by Longo and Calude https:…
Very large databases...and data analytics [are] a remarkable new field of investigation in computer science. The effectiveness of these tools is used to support a “philosophy” against the scientific method as developed throughout history. https://t.co/0zvK
@nntaleb Two references on this subject : "The deluge of spurious correlations in big data" by Longo and Calude https://t.co/jTtc3pTzhC … and the book "The AI delusion" by Gary Smith.
@randomuserbr @ScienceNews "... the more data, the more arbitrary, meaningless and useless (for future action) correlations will be found in them." https://t.co/b3b7M21vW9
@nntaleb Two bibliographical pointers on this very important subject : deluge of spurious correlations in big data by Longo and Calude https://t.co/jTtc3pTzhC and the book "The AI delusion" by Gary Smith.
RT @pp0196: @AedinCulhane @WiringTheBrain Mathematical/philosophical exposition https://t.co/hbkIQKfIQN Empirical estimation for specific d…
@AedinCulhane @WiringTheBrain Mathematical/philosophical exposition https://t.co/hbkIQKfIQN Empirical estimation for specific data (at least 25% of hard spurious correlation for a very). https://t.co/2bOqzOsfNl
@DrMJoyner @dgmacarthur @markmccarthyoxf @LordGenome @ewanbirney @MichelleNMeyer @nccomfort @StuartJRitchie @RAhlskog @kph3k @techreview @Graham_Coop There is a whole lot of shall i say bold statement regarding associations vs mechanism here. There is a wh
@eggersnsf Oberflächlich? https://t.co/41MjsWvBGL. Veraltet? Wenn veraltet gute Wissenschaft bedeutet, dann gern und lieber veraltet. Werden Patientinnen . . . früh entlassen? Ja, deswegen wurden die Fallpauschalen eingeführt. Ein gutes Beispiel für überfl
@petrodanylos @nikitakarachoi https://t.co/RbUrhPBFcd Produire des corrélations de cette magnitude n'est pas surprenant vu la taille des données. Ce qui est le important c'est que d'un GWAS à l'autres il y a très très peu de corrélations entres les scores
@MichelleNMeyer @itsbirdemic Big data are correlation mills. https://t.co/RbUrhPk4kF
@coherentstates Also spurious correlations exist in big data sets: https://t.co/2eNXaYR8QA
"The Deluge of Spurious Correlations in Big Data" "... we prove that very large databases have to contain arbitrary correlations." Meaning... Problems for GWAS and other large databases used for these types of association studies. https://t.co/4A7cRrkbK
A inevitabilidade das correlações espúrias que só surgem no volume de informações típico do Big Data https://t.co/BWh6q9ebzR
RT @WanderingWim: A sobering (and mathematical/philosophical) view of Big Data "science": The deluge of spurious correlations. https://t.co…
A sobering (and mathematical/philosophical) view of Big Data "science": The deluge of spurious correlations. https://t.co/e7LgMaBXsv
@Snowden Did you see this? The Deluge of Spurious Correlations in Big Data: https://t.co/EZ0WoT3QNN