Data Visualization and Data Science consultant.
Currently at UFRJ/PPGI Master's program in Computer Science (Data Visualization and Machine Learning).
function varianceBinarySearch(points, i, entropy, error=1e-6, steps=40) {
function recursiveBinarySearch(variance, lowerBound, upperBound) {
const probabilities = computeP_i(points, i, variance),
currentEntropy = computeEntropy(probabilities);
if (steps-- <= 1 || Math.abs(currentEntropy - entropy) <= error)
return probabilities;
if (currentEntropy > entropy)
return recursiveBinarySearch((lowerBound+variance)/2, lowerBound, variance);
else
return recursiveBinarySearch((variance+upperBound)/2, variance, upperBound);
}
const MINVAR = 0.01,
MAXVAR = 50;
return recursiveBinarySearch((MINVAR+MAXVAR)/2, MINVAR, MAXVAR);
}
function adjustVariances(points, entropy, error, steps) {
let size = points.length,
p = [];
for (let i = 0; i < size; i++)
p = p.concat(varianceBinarySearch(points, i, entropy, error, steps))
var total = p.reduce((sum, current)=>sum+current,0)
return p.map((x)=>x/total);
}
pointsMap = {
if (iteractionCount >= T)
yield yDisplay.map((p, i)=>(p.y==undefined?{x:Y[i][0],
name:p.name,
Cluster:p.Cluster}:
{x:Y[i][0],
y:Y[i][1],
name:p.name,
Cluster:p.Cluster}));
else {
// Adapted from: https://github.com/karpathy/tsnejs
var N = Y.length;
var dim = Y[0].length;
var grad = gradient(Y, pFinal, iteractionCount);
var sign = ((p)=>p<0?-1:p>0?1:p);
mutable costRun = cost(Y, pFinal);
mutable iteractionCount++;
// perform gradient step
for(var i=0;i<N;i++) {
for(var d=0;d<dim;d++) {
var gid = grad[i][d];
var sid = ystep[i][d];
var gainid = gains[i][d];
// compute gain update
var newgain = sign(gid) === sign(sid) ? gainid * 0.8 : gainid + 0.2;
if(gainid < 0.01) gainid = 0.01; // clamp
gains[i][d] = newgain; // store for next turn
// compute momentum step direction
var momval = momentum[iteractionCount < triggerMomentum]; // ? 0.5 : 0.8;
var newsid = momval * sid - eta * newgain * grad[i][d];
ystep[i][d] = newsid; // remember the step we took
// step!
Y[i][d] += newsid;
}
}
yield yDisplay.map((p, i)=>(p.y==undefined?{x:Y[i][0],
name:p.name,
Cluster:p.Cluster}:
{x:Y[i][0],
y:Y[i][1],
name:p.name,
Cluster:p.Cluster}));
}
}
Purpose-built for displays of data
Observable is your go-to platform for exploring data and creating expressive data visualizations. Use reactive JavaScript notebooks for prototyping and a collaborative canvas for visual data exploration and dashboard creation.
