RT @chris_bail: 1/n Can machine learning solve social science problems? This unprecedented study led by @msalganik and dozens of co-authors…
RT @chris_bail: 1/n Can machine learning solve social science problems? This unprecedented study led by @msalganik and dozens of co-authors…
this is fascinating!! CSS giant study on value of ML "only somewhat better than the results from a simple benchmark model that used linear regression (continuous outcomes) or logistic regression (binary outcomes) with four predictor variables selected by a
RT @autismcrisis: "Hundreds of researchers attempted to predict six life outcomes" but "no one made very accurate predictions"? https://t.c…
Also kudos to the massive collaboration!
This is a fascinating study. Especially for those interested in predictive modelling in the social sciences.
RT @chris_bail: 1/n Can machine learning solve social science problems? This unprecedented study led by @msalganik and dozens of co-authors…
RT @chris_bail: 1/n Can machine learning solve social science problems? This unprecedented study led by @msalganik and dozens of co-authors…
RT @chris_bail: 1/n Can machine learning solve social science problems? This unprecedented study led by @msalganik and dozens of co-authors…
RT @chris_bail: 1/n Can machine learning solve social science problems? This unprecedented study led by @msalganik and dozens of co-authors…
RT @chris_bail: 1/n Can machine learning solve social science problems? This unprecedented study led by @msalganik and dozens of co-authors…
"Hundreds of researchers attempted to predict six life outcomes" but "no one made very accurate predictions"? https://t.co/tPsM7zoFCu in PNAS, autism-relevant, free
1/n Can machine learning solve social science problems? This unprecedented study led by @msalganik and dozens of co-authors is a #kaggle-style competition to predict outcomes of children from vulnerable families using held out data from a multiwave study:
RT @BillMillerMD: Measuring the predictability of life outcomes with a scientific mass collaboration | PNAS https://t.co/OlHGfx3iBf
Measuring the predictability of life outcomes with a scientific mass collaboration | PNAS https://t.co/OlHGfx3iBf
RT @eredmil1: Great thread supporting why often "simple" regression is the best way to explore social scientific questions around IV effect…
RT @ChrisDanforth: Predicting life outcomes is hard: "Despite using a rich dataset & machine-learning methods... the best predictions were…
RT @eredmil1: Great thread supporting why often "simple" regression is the best way to explore social scientific questions around IV effect…
RT @ChrisDanforth: Predicting life outcomes is hard: "Despite using a rich dataset & machine-learning methods... the best predictions were…
Measuring the predictability of life outcomes with a scientific mass collaboration https://t.co/2qrtZSSKam
RT @ChrisDanforth: Predicting life outcomes is hard: "Despite using a rich dataset & machine-learning methods... the best predictions were…
The Fragile Families (@FFCWS) paper is out . If you heard my talk about ethics in AI at @RSAConference with @BDayanim this year, we referenced it. (Upshot: can algorithms be used to predict wellbeing.) Spoiler alert: no. https://t.co/jgALv3Txm8 https:/
Predicting life outcomes is hard: "Despite using a rich dataset & machine-learning methods... the best predictions were not very accurate, only slightly better than a simple benchmark model." Fragile Families Challenge, huge effort by many teams htt
Great thread supporting why often "simple" regression is the best way to explore social scientific questions around IV effects on a DV. ML/prediction may be hot, but they aren't always the right way to answer your scientific question.
RT @LandonSchnabel: Excited to see that the Fragile Families Challenge paper is now online! https://t.co/716HRhbGAM While social science c…
RT @BeEngelhardt: Extremely excited to see the Fragile Families Challenge paper out. @msalganik asked if ML worked to predict life outcomes…
RT @BeEngelhardt: Extremely excited to see the Fragile Families Challenge paper out. @msalganik asked if ML worked to predict life outcomes…
RT @BeEngelhardt: Extremely excited to see the Fragile Families Challenge paper out. @msalganik asked if ML worked to predict life outcomes…
Excited to see that the Fragile Families Challenge paper is now online! https://t.co/716HRhbGAM While social science can predict life outcomes, these outcomes are also shaped by things we don't measure—and, in certain ways, are perhaps simply idiosyncrati
Extremely excited to see the Fragile Families Challenge paper out. @msalganik asked if ML worked to predict life outcomes with complex small sample size data. He & @IanLundberg1 built the train & test sets from @FFCWS data & let everyone take a
AI and Life trajectories
RT @DanielJDrucker: #MachineLearning hits a reality bump. “Additionally, researchers must reconcile the idea that they understand life tra…
#MachineLearning hits a reality bump. “Additionally, researchers must reconcile the idea that they understand life trajectories with the fact that none of the predictions were very accurate” @PNASNews https://t.co/q65XX5yr3k