The Least Squares Projection

Beneath the graph are the data for the problem. You have three observations on two independent variables, x. The black directed line segments with blue balls represent the vectors in x. There are three observations on a dependent variable, y. The blue directed line segment represents the vector y. You can change any of the data points. The 2 independent variables span a two dimensional vector space, represented by the plane. The least squares combination of the independent variables is represented by the red directed line segment in the plane. The orange directed line segment is the least squares residual vector. Regardless of the data values you choose, the space spanned by the columns of x will remain orthogonal to the error vector. Try changing the data points. Observe how the plane and vectors are reoriented. The least squares coefficients are also recomputed. You can also left click and drag the cursor across the graph to get a different perspective.

  No support for LM Objects