least squares
Minimize residual errors
Least squares is a mathematical method for finding the best-fitting curve or line through a set of data points by minimizing the sum of the squared differences between observed values and predicted values.
The technique is fundamental to regression analysis and statistical modeling: We used least squares to fit a trend line to the sales data.
Ongoing compensation for rebroadcasts
- coefficientNumber multiplying a variable
- interceptSeize en route
- predictorsThings that forecast outcomes
- dependent variableOutcome being measured
independent variables · outliers · leverage · influence · confidence intervals · diagnostics · normality
Approximating data mathematically
- data pointsDiscrete measured values
- trendlineStatistical direction indicator
- scatter plotGraph showing paired data points
- splineFlexible strip for drafting
interpolation · extrapolation · calibration · overfitting · underfitting · degrees of freedom · smoothing
Making as small as possible
- convergenceProcess of approaching sameness
- iterativeInvolving repeated cycles
- convexityOutward-curving quality
- local minimumLowest nearby, not overall
Ordered numerical arrangement
- projectionSomething that extends outward
- orthogonalityState of being independent
- QR decompositionMatrix as orthogonal-triangular product
- rankPosition in hierarchy
Sample-based inference rule
- unbiasedFree from prejudice
- varianceSpread around the mean
- standard errorStatistic's expected deviation measure
- hypothesis testingStatistical method for evaluating claims
F-test · t-test · p-value · ANOVA · degrees of freedom · consistency · efficiency
residuals
Ongoing compensation for rebroadcasts
- coefficientNumber multiplying a variable
- interceptSeize en route
- predictorsThings that forecast outcomes
dependent variable · independent variables · outliers · leverage · influence · confidence intervals · diagnostics · normality
curve fitting
Approximating data mathematically
- data pointsDiscrete measured values
- trendlineStatistical direction indicator
- scatter plotGraph showing paired data points
spline · interpolation · extrapolation · calibration · overfitting · underfitting · degrees of freedom · smoothing
minimization
Making as small as possible
- convergenceProcess of approaching sameness
- iterativeInvolving repeated cycles
- convexityOutward-curving quality
local minimum · Lagrange multipliers · regularization · constraint
matrix
Ordered numerical arrangement
- projectionSomething that extends outward
- orthogonalityState of being independent
- QR decompositionMatrix as orthogonal-triangular product
estimator
Sample-based inference rule
- unbiasedFree from prejudice
- varianceSpread around the mean
- standard errorStatistic's expected deviation measure
hypothesis testing · F-test · t-test · p-value · ANOVA · degrees of freedom · consistency · efficiency
- linear algebramatrix, linear equation, vectors, linear function, matrices
- machine learningneural networks, deep learning, backpropagation, underfitting, CNNs
- approximationinterpolation, spline interpolation, extrapolation, curve fitting
- computational mathalgorithm, linear algebra, calculus, differential equations, finite element
- numerical methodsbisection, interpolation, Gaussian elimination, convergence, approximation
- regressionlinear regression, residuals, residual, MARS