Python scipy.stats.ranksums() Examples

def rank_test(covariates, groups):
    """ 
    Wilcoxon rank sum test for the distribution of treatment and control covariates.
    
    Parameters
    ----------
    covariates : DataFrame 
        Dataframe with one covariate per column.
        If matches are with replacement, then duplicates should be 
        included as additional rows.
    groups : array-like
        treatment assignments, must be 2 groups
    
    Returns
    -------
    A list of p-values, one for each column in covariates
    """    
    colnames = list(covariates.columns)
    J = len(colnames)
    pvalues = np.zeros(J)
    for j in range(J):
        var = covariates[colnames[j]]
        res = ranksums(var[groups == 1], var[groups == 0])
        pvalues[j] = res.pvalue
    return pvalues 

def compare_by_position(bed1,bed2,xmfa):
    pos_dict = {}

    for i,bed in enumerate([bed1,bed2]):
        pos_dict[i] = {}
        with open(bed,'r') as fi:
                for line in fi:
                #2  1892198 1892199 TCMMTMTTMMM 0.5 -   16
                    csome,start,end,motif,perc_meth,strand,num_reads,probabilities = tuple(line.split('\t'))
                    pos_dict[i][(csome,start,end,strand)] = ((perc_meth,num_reads),np.asarray([float(p) for p in probabilities.strip().split(',')]))

    for pos in pos_dict[0]:
        if pos in pos_dict[1]:
            try:
                u,pval = mannwhitneyu(pos_dict[0][pos][1],pos_dict[0][pos][1],alternative='two-sided')
            except ValueError:
                u,pval = 'none','identical'
            u2,pval2 = ranksums(pos_dict[0][pos][1],pos_dict[0][pos][1])
            try:
                t,pval3 = ttest_ind(pos_dict[0][pos][1],pos_dict[0][pos][1])
            except:
                t,pval3 = 'none','missing df'
            d,pval4 = ks_2samp(pos_dict[0][pos][1],pos_dict[0][pos][1])
            if pval4 < 0.9:
                print pos, pos_dict[0][pos][0], pos_dict[1][pos][0], pval, pval2, pval3, pval4 

def test_ranksums_result_attributes(self):
        res = stats.ranksums(np.arange(5), np.arange(25))
        attributes = ('statistic', 'pvalue')
        check_named_results(res, attributes) 

def getParent(self,A,TP):
		#A is the cluster,we are trying to find the parent for cluster A.
		#TP is the parent level
		#PL is the cluste list in parent level
	
		PL=[item for item in self.Nodes if (item.T==TP and item.ST<=A.ST)]
		PL=sorted(PL,key=lambda item:sum([getDistance(item,ck) for ck in A.cells])/len(A.cells))
		#pdb.set_trace()
		#---------------------------------------------------------------
		# if time sync is disabled, output the closest node in parent level
		
		if self.dsync=='True' or self.dsync=='1':
			if len(PL)>0:
				return PL[0]
			return None
			
		#--------------------------------------------------------------
		
		pvcut=0.1
		if len(PL)>1:
			X=[getDistance(PL[0],item) for item in A.cells]
			Y=[getDistance(PL[-1],item)  for item in A.cells]
			# Length adjustment if Vector is too short
			SizeFactor=50 # Used for length adjustment
			X=X*int(SizeFactor/len(X)) if len(X)<SizeFactor else X
			Y=Y*int(SizeFactor/len(Y)) if len(Y)<SizeFactor else Y
			pv=ranksums(X,Y)[-1]
			if pv<pvcut:
				return PL[0]
			else:
			   return self.getParent(A,TP-1)
		elif len(PL)==1:
			if PL[0]==self.Nodes[0]:
				return PL[0]
			else:
				PLL=[item for item in self.Nodes if (item.T==TP-1 and item.ST<=A.ST)]
				if len(PLL)==1:
					return PL[0]
				else:
					return self.getParent(A,TP-1)
		else:
			if A!=self.Nodes[0]:
				return self.getParent(A,TP-1) 

def ranksum_test(self, metric_name=None):
        """
        Non-parametric test for testing consistent differences between pairs
        of obeservations.
        The test counts the number of observations that are greater, smaller
        and equal to the mean
        `<http://en.wikipedia.org/wiki/Wilcoxon_rank-sum_test>`_.
        """
        self._check_is_evaluated()
        metric_name = self._validate_metric_name(metric_name)
        metrics_per_estimator_dataset = \
            self._get_metrics_per_estimator_dataset(
                metric_name)

        ranksum_df = pd.DataFrame()
        perms = itertools.product(metrics_per_estimator_dataset.keys(),
                                  repeat=2)
        values = np.array([])
        for perm in perms:
            x = metrics_per_estimator_dataset[perm[0]]
            y = metrics_per_estimator_dataset[perm[1]]
            t_stat, p_val = ranksums(x, y)
            ranksum = {
                "estimator_1": perm[0],
                "estimator_2": perm[1],
                "t_stat": t_stat,
                "p_val": p_val
            }
            ranksum_df = ranksum_df.append(ranksum, ignore_index=True)
            values = np.append(values, t_stat)
            values = np.append(values, p_val)

        index = ranksum_df["estimator_1"].unique()
        values_names = ["t_stat", "p_val"]
        col_idx = pd.MultiIndex.from_product([index, values_names])
        values_reshaped = values.reshape(len(index),
                                         len(values_names) * len(index))

        values_df_multiindex = pd.DataFrame(values_reshaped, index=index,
                                            columns=col_idx)

        return ranksum_df, values_df_multiindex