Function pearson_log_likelihood_ratio

pattern/metrics.py:788–805

Returns (G, P) for the n x m observed and expected data (containing absolute frequencies). If expected is None, an equal distribution over all classes is assumed. If df is None, it is the number of classes-1 (i.e., len(observed[0]) - 1). P < 0.05: significant P < 0.0

(observed=[], expected=[], df=None, tail=UPPER)

#--- PEARSON'S LOG LIKELIHOOD RATIO APPROXIMATION --------------------------------------------------

def pearson_log_likelihood_ratio(observed=[], expected=[], df=None, tail=UPPER):
    """ Returns (G, P) for the n x m observed and expected data (containing absolute frequencies).
        If expected is None, an equal distribution over all classes is assumed.
        If df is None, it is the number of classes-1 (i.e., len(observed[0]) - 1).
        P < 0.05: significant
        P < 0.01: very significant
    """
    O  = observed
    E  = expected or _expected(observed)
    df = df or (len(O) > 0 and len(O[0])-1 or 0)
    G  = 0.0
    for i in range(len(O)):
        for j in range(len(O[i])):
            if O[i][j] != 0 and E[i][j] != 0:
                G += O[i][j] * log(O[i][j] / E[i][j])
    G = G * 2
    P = gammai(df * 0.5, G * 0.5, tail)
    return (G, P)
    
llr = likelihood = pearson_log_likelihood_ratio