Plotting TemplateMarkdownTeXArrays and ObjectsPlotting Method 1Plotting Method 2Plotting Method 3Plotting ExampleHorizontal LinesCars/Horizontal LineSlopeLinesBest Fitting LineParabolaSineExponential FunctionLogarithmic FunctionGrowth RatesSigmoid FunctionDerivative of ParabolaDerivative of SineDerivative of Exponential FunctionDerivative Using Chain RuleBlocks of CodeFor LoopsGradient DescentGradient DescentThe MSEMSE With Just SlopeGradient Descent ProgramClassesRegression ModulePenguinsPenguins ApplicationBaseball ApplicationPlanesParaboloidSine WaveAnother Sine WaveThe GradientGradient Descent with ParaboloidGradient Descent with a Sine WaveGradient Descent with another Sine WaveRegression RevisitedMinimizing MSERegression ClassPenguinsDerivative PracticeExam Practice
Regression Class (with second predictor)
class Regression{
constructor(X1,X2,Y,rate,epochs){
this.X1 = X1
this.X2 = X2
this.Y = Y
this.rate = rate
this.epochs = epochs
}
mse(w2,w1,b){
const Yhat =this.X1.map((x,i) => w2*this.X2[i] + w1*this.X1[i] + b)
const differences = this.X1.map((x,i)=>this.Y[i]-Yhat[i])
const squares = differences.map(d => d*d)
const mse = d3.mean(squares)
return mse
}
msePartialw2(w2,w1,b){
const Yhat =this.X1.map((x,i) => w2*this.X2[i] + w1*this.X1[i] + b)
const summand = this.X1.map((x,i) => (this.Y[i]-Yhat[i])*this.X2[i])
return -2*d3.sum(summand)/this.X1.length
}
msePartialw1(w2,w1,b){
const Yhat =this.X1.map((x,i) => w2*this.X2[i] + w1*this.X1[i] + b)
const summand = this.X1.map((x,i) => (this.Y[i]-Yhat[i])*this.X1[i])
return -2*d3.sum(summand)/this.X1.length
}
msePartialb(w2,w1,b){
const Yhat =this.X1.map((x,i) => w2*this.X2[i] + w1*this.X1[i] + b)
const summand = this.X1.map((x,i) => (this.Y[i]-Yhat[i]))
return -2*d3.sum(summand)/this.X1.length
}
fit(){
let w1 = 0
let w2 = 0
let b = 0
for(let i = 0; i <= this.epochs; i = i + 1){
w1 = w1 - this.rate*this.msePartialw1(w2,w1,b)
w2 = w2 - this.rate*this.msePartialw2(w2,w1,b)
b = b - this.rate*this.msePartialb(w2,w1,b)
}
return [w2,w1,b]
}
}
X1 = [3,6,4,1]
X2 = [7,6,9,5]
Y = [3,7,5,8]
regression = new Regression(X1,X2,Y,.0001,1000)
p = regression.fit()
regression.mse(p[0],p[1],p[2])
One platform to build and deploy the best data apps
Experiment and prototype by building visualizations in live JavaScript notebooks. Collaborate with your team and decide which concepts to build out.
Use Observable Framework to build data apps locally. Use data loaders to build in any language or library, including Python, SQL, and R.
Seamlessly deploy to Observable. Test before you ship, use automatic deploy-on-commit, and ensure your projects are always up-to-date.