Visualizzazione post con etichetta statistica e rischio. Mostra tutti i post
Visualizzazione post con etichetta statistica e rischio. Mostra tutti i post

giovedì 3 maggio 2018

BAYES VS REGRESSIONI

RRiccardo Mariani
Adesso


BAYES VS REGRESSIONI


La differenza si vede nei metodi di programmazione, nel caso si utilizzi la regressione basterà inserire la funzione teorica ricavata nel software in forma di algoritmo, nel caso si utilizzi Bayes bisogna ricorrere ad un algoritmo che si aggiorni continuamente sulla base dei successi e degli insuccessi pratici che si riscontrano. La regressione fornisce soluzioni purché stiano su una funzione lineare (a volte una retta), l'algoritmo bayesiano non ha questo vincolo, valorizza ogni dato che incontra (tutto c'entra con tutto) e si rende disponibile grazie alla potenza di calcolo delle nuove macchine.

AMAZON.COM
"What does AI mean for your business? Read this book to find out." -- Hal Varian, Chief Economist, GoogleArtificial intelligence does the seemingly impossible, magically bringing machines to life--driving cars, trading stocks, and teaching children. But facing the sea change that AI wil...
iccardo Mariani
Adesso

BAYES VS REGRESSIONI
La differenza si vede nei metodi di programmazione, nel caso si utilizzi la regressione basterà inserire la funzione teorica ricavata nel software in forma di algoritmo, nel caso si utilizzi Bayes bisogna ricorrere ad un algoritmo che si aggiorni continuamente sulla base dei successi e degli insuccessi pratici che si riscontrano.

AMAZON.COM
"What does AI mean for your business? Read this book to find out." -- Hal Varian, Chief Economist, GoogleArtificial intelligence does the seemingly impossible, magically bringing machines to life--driving cars, trading stocks, and teaching children. But facing the sea change that AI wil...

mercoledì 16 marzo 2016

Fifteen Matters of Life and Death - More Sex Is Safer Sex: The Unconventional Wisdom of Economics by Steven E. Landsburg

Fifteen Matters of Life and Death - More Sex Is Safer Sex: The Unconventional Wisdom of Economics by Steven E. Landsburg - #ventilatoripertutti #schellingavermicino #valoredellavitaevaloririschiosi #tuttiigiorniprezziamolavita #cellulariallaguida #blackstonetroppoprodigo #fareunastimadadoveiniziare
Fifteen Matters of Life and DeathRead more at location 2540
Note: Aiutare in modo empatico non significa dare cose anziché denaro ma "dare". Posso aiutare l'accaldato dandogli un ventilatore o dei soldi (taglio tasse). Solo la seconda soluzione è rispettosa delle volontà del prossimo Quel che non si vede: alfredino e il paracari di Shelling. La tassazione empatica come tassazione irrazionale Vite monetizzate. La tassa x scampare King Kong Edit
Note: 15@@@@@@@@@@@@@ Edit
There is room for a great deal of disagreement about how much assistance rich people should give to poor people, either voluntarily or through the tax system. But surely whatever we do spend should be spent in the ways that are most helpful. Therefore there’s no use arguing that the real trade-off should not be ventilators versus milk but ventilators versus tax cuts, or ventilators versus foreign wars.Read more at location 2565
Note: VENTILATORI PER TUTTI! Edit
The bloggers at Daily Kos would have us guarantee ventilator support to everyone who can’t afford it. For the same cost, we could give each of those people a choice between ventilator insurance or $75 cash.Read more at location 2569
Note: CASH O VENTILATORI? Edit
Remarkably, Professor Frank admits that most poor people would prefer the milk and eggs, but still argues for the ventilators on the grounds that it would make the rest of us feel better!Read more at location 2583
Note: VENTILARE LA COSCIENZA DEL DONATORE Edit
But ignoring other people’s needs to make yourself feel better is the very opposite of sympathy and empathy.Read more at location 2584
Many years ago, the Nobel laureate Thomas Schelling asked why communities will sometimes spend millions to save the life of a known victim—say, a trapped miner—but refuse to spend even, say, $200,000 on a highway guardrail that would save an average of one life per year. His answer was to distinguish between “identified lives”—like the trapped miner, or Tirhas Habtegiris—and “statistical lives,” like the unknown beneficiaries of the guardrail. Robert Frank embraces this distinction. For some reason, we are supposed to care more about identified lives than statistical lives,Read more at location 2587
Note: IL PARACARRO DI SCHELLING Edit
Besides being incoherent, the distinction between a statistical and an identified life is morally obtuse.Read more at location 2601
The identified/statistical life approach says, in essence, that I should prefer three deaths to one death as long as the three victims happen to be invisible.Read more at location 2613
Note: VISIBILE INVISIBILE Edit
If I could save you $5 in taxes by releasing a toxic chemical that might kill you—say with a probability of one in a million—I wouldn’t do it. If I could save you $20 in taxes by releasing the same chemical, I probably would.Read more at location 2616
Note: VALORE DELLA VITA Edit
I know this because economists have made a habit of observing people’s choices—for example, the size of the pay cuts they’ll accept to move into safer jobs. On that basis, Harvard Law professor Kip Viscusi estimates that the average American will pay about $5 to steer clear of a one-in-a-million chance of being killed. For blue-collar women, it’s closer to $7, and for blue-collar men it’s even higher.Read more at location 2620
Note: LAVORI RISCHIOSI Edit
Your life is probably worth somewhere between five and ten million dollars.Read more at location 2628
If King Kong is likely to kill 300 Americans out of 300 million, there’s a one in a million chance you’ll be a victim. If we can stop him with $300 million worth of ape repellent—that is, $1 million per life—your share of the tax bill comes to about a dollar. That’s a good deal. But if instead it costs $30 billion to stop Kong, then it’s better to let him wreak havoc. Your share of that $30 billion would be $100, and we know from Professor Viscusi’s studies that almost nobody wants to pay that much to avoid a one-in-a-million chance of death.Read more at location 2631
Note: FERMARE KING KONG Edit
Cell-phone use increases your accident risk by almost 400 percent. But so what?Read more at location 2643
Note: CELLULARE ALLA GUIDA Edit
Just getting behind the wheel (as opposed to staying home in bed) multiplies your accident risk by far more than 400 percent, but so far the Tappets have not proposed to outlaw driving.Read more at location 2645
Note: INCIDENTI Edit
The analysis is courtesy of Brookings Institution economists Robert Hahn, Paul Tetlock, and Jason Burnett, who reckon that drivers’ cell phones are indeed deadly but nevertheless (on balance) a good thing.Read more at location 2651
The researchers estimate that in one year, driver use of cell phones caused about 300 fatalities, 38,000 nonfatal injuries, and 200,000 damaged vehicles.Read more at location 2657
the price is $6.6 million, a widely used standard based on Kip Viscusi’s analysis.Read more at location 2662
You implicitly put a dollar value on human lives every time you buy a candy bar with funds that could instead have been donated to the local fire department. No matter who you are, there is a limit to what you’re willing to spend to save lives;Read more at location 2664
Note: OGNI GIORNO QUANTIFICHIAMO LA VITA Edit
the only question is whether you’re willing to think honestly about what that limit is.Read more at location 2666
Note: ONESTÀ Edit
Pricing the fatalities at $6.6 million each, and adding in the costs of injuries and vehicle damage, Mr. Hahn and his colleagues estimate that in one year, cell-phone use by drivers caused $4.6 billion worth of damage.Read more at location 2669
the value of a cell-phone call is equal to what you’re willing to pay for it, minus what you actually pay for it. Willingness to pay is estimated from demand studies (factoring in the truism that some calls are worth more than others). Actual charges, of course, come from real-life cell-phone bills.Read more at location 2672
Note: VALORE DELLA CHIAMATA: IL PIACRRE CHE MI DÀ (QUANTO SAREI DISPOSTO A PAGARE) MENO QUANTO PAGO. Edit
in one year, cell-phone calls made by drivers had a total value of $25 billion. That $25 billion benefit beats the $4.6 billion cost,Read more at location 2675
Actually, I don’t buy it, for a variety of reasons. First, drivers make a lot of calls that could easily wait for the next rest stop.Read more at location 2676
Note: DUBBI Edit
Moreover, Hahn et al missed a potentially important factor on the cost side: they counted fatalities, they counted injuries, they counted property damage, but they failed to account for the inconvenience to people who choose not to drive because cell phones have made driving more dangerous.Read more at location 2680
Note: E CHI RINUNCIA A GUIDARE? Edit
Over two centuries ago, a lawyer named William Blackstone declared that it’s better for ten guilty persons to escape than for one innocent to suffer. Why ten, as opposed to, say, twelve or eight?Read more at location 2691
Note: LO STANDARD DI BLACKSTONE Edit
for two centuries, legal scholars have cited Blackstone’s refusal to think and mistaken it for an example of a thought.Read more at location 2693
Note: LIQUIDARE IL TRADE OFF Edit
The cost (to you) of a false conviction is that you might be the unlucky innocent who goes to jail. The cost (to you) of a false acquittal is that you might cross paths with the criminal we just freed (or with some other criminal who’s feeling emboldened by all these false acquittals). A “ten guilty men” standard saddles you with one bundle of risks; a “five guilty men” or a “hundred guilty men” standard saddles you with another. The right standard is the one that saddles you with the burden you prefer (where “prefer” means “dislike the least”).Read more at location 2699
Note: UNO STANDARD EQUO Edit
Ten false acquittals in murder cases give you ten chances to cross paths with a newly released murderer. One false conviction gives you one chance of landing (at worst) in the electric chair. I’m not so fond of either prospect, but given a choice, I’ll take the latter.Read more at location 2704
Note: BLACKSTONE TROPPO PRODIGO? Edit
I expect we’d all be happier with, say, a “three guilty men” standard: if we’re 75 percent sure you did the crime, you do the time.Read more at location 2706
The cutoff for reasonable doubt should be at the odds we find just barely acceptable.Read more at location 2709
Note: RAGIONEVOLE DUBBIO Edit
If you never miss a plane, you’re spending too much time in airports; if you never convict an innocent, you’re not convicting enough of the guilty.Read more at location 2715
Note: TRADE OFF Edit
But I am substantially less comfortable with it in a world where policemen (and others) sometimes manufacture evidence against peopleRead more at location 2717
Note: POLIZIA CORROTTA Edit
That makes me want to nudge the cutoff up a bit, probably to something very like Blackstone’s ninety-something-percent standard.Read more at location 2718
If we execute murderers, why not execute the people who write computer worms? According to the math on the back of my envelope, it would be a better investment.Read more at location 2720
Note: PENA DI MORTE X CHI INFETTA I PC Edit
First, what’s the value of executing a murderer? A high-end estimate is that each execution deters about ten murdersRead more at location 2722
Note: I HENEGIVI DI UCCIDERE UN ASSASSINO Edit
So call it ten lives saved, with a value—again a high-end estimate—of about $10 million apiece. Then the benefit of executing a murderer is roughly ten times $10 million, or $100 million—andRead more at location 2723
It’s estimated that vermiscripting and related activities cost the world about $50 billion a year.Read more at location 2727
There are a lot of reasons to believe the number is seriously overblown, not least because victimized firms tend to exaggerate the damage when they’re making insurance claims.Read more at location 2729
Given that $50 billion figure, all we’d have to do is deter just one-fifth of one percent of all vermiscripting for just one year in order to gain the same $100 million benefit we get from executing a killer.Read more at location 2731
Executing the murderer means giving you the safety. Executing the vermiscriptor means giving you the cash. You’d rather have the cash than the safety. Ergo, executing the vermiscriptor is the better policy.Read more at location 2755
Note: MEGLIO IL CASH CHE L ASSICURAZIONE Edit
There’s at least one exceptionRead more at location 2757
Note: ALTRRNATIVA Edit
If we can effectively deter malicious hackers by cutting off their supply of Twinkies or crippling their EverQuest avatars, then there’s no need to fry them.