fix: PointSetDist_Scoring.WithScalarAnswer.scoreWithPrior

Done in pair coding with Quinn.
Value::[0.3 to 0.9]
This commit is contained in:
NunoSempere 2022-06-20 16:14:41 -04:00
parent 7e859c7823
commit 4b1c226173
2 changed files with 91 additions and 1 deletions

View File

@ -0,0 +1,81 @@
// Bring up a discrete distribution
open Jest
open Expect
open TestHelpers
open GenericDist_Fixtures
// WithDistAnswer -> in the KL divergence test file.
// WithScalarAnswer
describe("WithScalarAnswer: discrete -> discrete -> float", () => {
let mixture = a => DistributionTypes.DistributionOperation.Mixture(a)
let pointA = mkDelta(3.0)
let pointB = mkDelta(2.0)
let pointC = mkDelta(1.0)
let pointD = mkDelta(0.0)
test("score: agrees with analytical answer when finite", () => {
let prediction' = [(pointA, 0.25), (pointB, 0.25), (pointC, 0.25), (pointD, 0.25)]->mixture->run
let prediction = switch prediction' {
| Dist(PointSet(a'')) => a''
| _ => raise(MixtureFailed)
}
let answer = 2.0 // So this is: assigning 100% probability to 2.0
let result = PointSetDist_Scoring.WithScalarAnswer.score(~estimate=prediction, ~answer)
switch result {
| Ok(x) => x->expect->toEqual(-.Js.Math.log(0.25 /. 1.0))
| _ => raise(MixtureFailed)
}
})
test("score: agrees with analytical answer when finite", () => {
let prediction' = [(pointA, 0.75), (pointB, 0.25)]->mixture->run
let prediction = switch prediction' {
| Dist(PointSet(a'')) => a''
| _ => raise(MixtureFailed)
}
let answer = 3.0 // So this is: assigning 100% probability to 2.0
let result = PointSetDist_Scoring.WithScalarAnswer.score(~estimate=prediction, ~answer)
switch result {
| Ok(x) => x->expect->toEqual(-.Js.Math.log(0.75 /. 1.0))
| _ => raise(MixtureFailed)
}
})
test("scoreWithPrior: ", () => {
let prior' = [(pointA, 0.5), (pointB, 0.5)]->mixture->run
let prediction' = [(pointA, 0.75), (pointB, 0.25)]->mixture->run
let prediction = switch prediction' {
| Dist(PointSet(a'')) => a''
| _ => raise(MixtureFailed)
}
let prior = switch prior' {
| Dist(PointSet(a'')) => a''
| _ => raise(MixtureFailed)
}
let answer = 3.0 // So this is: assigning 100% probability to 2.0
let result = PointSetDist_Scoring.WithScalarAnswer.scoreWithPrior(
~estimate=prediction,
~answer,
~prior,
)
switch result {
| Ok(x) => x->expect->toEqual(-.Js.Math.log(0.75 /. 1.0) -. -.Js.Math.log(0.5 /. 1.0))
| _ => raise(MixtureFailed)
}
})
})
// WithDistAnswer
/*
describe("WithScalarAnswer: discrete -> discrete -> float", () => {
})
// TwoScalars
describe("WithScalarAnswer: discrete -> discrete -> float", () => {
})
*/

View File

@ -115,11 +115,20 @@ module WithScalarAnswer = {
} }
_score(~estimatePdf, ~answer) _score(~estimatePdf, ~answer)
} }
/*
let score1 = (~estimate: pointSetDist, ~answer: scalar): result<score, Operation.Error.t> => {
let probabilityAssignedToAnswer = Ok(1.0)
}
*/
let scoreWithPrior = (~estimate: pointSetDist, ~answer: scalar, ~prior: pointSetDist): result< let scoreWithPrior = (~estimate: pointSetDist, ~answer: scalar, ~prior: pointSetDist): result<
score, score,
Operation.Error.t, Operation.Error.t,
> => { > => {
E.R.merge(score(~estimate, ~answer), score(~estimate=prior, ~answer))->E.R2.fmap(((s1, s2)) =>
s1 -. s2
)
/*
let _scoreWithPrior = ( let _scoreWithPrior = (
~estimatePdf: float => float, ~estimatePdf: float => float,
~answer: scalar, ~answer: scalar,
@ -132,7 +141,6 @@ module WithScalarAnswer = {
} else if numerator == 0.0 || priorDensityOfAnswer == 0.0 { } else if numerator == 0.0 || priorDensityOfAnswer == 0.0 {
infinity->Ok infinity->Ok
} else { } else {
minusScaledLogOfQuotient(~esti=numerator, ~answ=priorDensityOfAnswer)
} }
} }
@ -149,6 +157,7 @@ module WithScalarAnswer = {
| Mixed(prio) => Mixed.T.xToY(x, prio)->sum | Mixed(prio) => Mixed.T.xToY(x, prio)->sum
} }
_scoreWithPrior(~estimatePdf, ~answer, ~priorPdf) _scoreWithPrior(~estimatePdf, ~answer, ~priorPdf)
*/
} }
} }