Please enable JavaScript.

Coggle requires JavaScript to display documents.

Supervised Learning (Terminology (error functional (is a quality…

- - - - How does the set of algorithms look like
        
        a(x)='sum from j=1 to j=d+1'(wj*xj)
        
        a(x)=<w,x> - the same as above but in a form of scalar product
      - what to use as an error functional
        
        |a(x)-y|
        
        (a(x)^2-y)
        
        Q(a,x)=1/l*'sum_from 1 to l'(a(xi)-yi)^2
      - Transition into the matrix form
        
        X is a training set and has objects-features
        
        ij element describes value of feature j for object i
        
        y is a vector of true values for each object
        
        Functional can be rewritten as a function Q(w,X)=1/l*||Xw-y||^2->min