Python statsmodels.stats.multitest.fdrcorrection() Examples

The following are 4 code examples of statsmodels.stats.multitest.fdrcorrection(). You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may also want to check out all available functions/classes of the module statsmodels.stats.multitest , or try the search function .
Example #1
Source File: multicomp.py    From vnpy_crypto with MIT License 6 votes vote down vote up 
def mcfdr(nrepl=100, nobs=50, ntests=10, ntrue=6, mu=0.5, alpha=0.05, rho=0.):
    '''MonteCarlo to test fdrcorrection
    '''
    nfalse = ntests - ntrue
    locs = np.array([0.]*ntrue + [mu]*(ntests - ntrue))
    results = []
    for i in range(nrepl):
        #rvs = locs + stats.norm.rvs(size=(nobs, ntests))
        rvs = locs + randmvn(rho, size=(nobs, ntests))
        tt, tpval = stats.ttest_1samp(rvs, 0)
        res = fdrcorrection_bak(np.abs(tpval), alpha=alpha, method='i')
        res0 = fdrcorrection0(np.abs(tpval), alpha=alpha)
        #res and res0 give the same results
        results.append([np.sum(res[:ntrue]), np.sum(res[ntrue:])] +
                       [np.sum(res0[:ntrue]), np.sum(res0[ntrue:])] +
                       res.tolist() +
                       np.sort(tpval).tolist() +
                       [np.sum(tpval[:ntrue]<alpha),
                        np.sum(tpval[ntrue:]<alpha)] +
                       [np.sum(tpval[:ntrue]<alpha/ntests),
                        np.sum(tpval[ntrue:]<alpha/ntests)])
    return np.array(results)
Example #2
Source File: multicomp.py    From Splunking-Crime with GNU Affero General Public License v3.0 6 votes vote down vote up 
def mcfdr(nrepl=100, nobs=50, ntests=10, ntrue=6, mu=0.5, alpha=0.05, rho=0.):
    '''MonteCarlo to test fdrcorrection
    '''
    nfalse = ntests - ntrue
    locs = np.array([0.]*ntrue + [mu]*(ntests - ntrue))
    results = []
    for i in range(nrepl):
        #rvs = locs + stats.norm.rvs(size=(nobs, ntests))
        rvs = locs + randmvn(rho, size=(nobs, ntests))
        tt, tpval = stats.ttest_1samp(rvs, 0)
        res = fdrcorrection_bak(np.abs(tpval), alpha=alpha, method='i')
        res0 = fdrcorrection0(np.abs(tpval), alpha=alpha)
        #res and res0 give the same results
        results.append([np.sum(res[:ntrue]), np.sum(res[ntrue:])] +
                       [np.sum(res0[:ntrue]), np.sum(res0[ntrue:])] +
                       res.tolist() +
                       np.sort(tpval).tolist() +
                       [np.sum(tpval[:ntrue]<alpha),
                        np.sum(tpval[ntrue:]<alpha)] +
                       [np.sum(tpval[:ntrue]<alpha/ntests),
                        np.sum(tpval[ntrue:]<alpha/ntests)])
    return np.array(results)
Example #3
Source File: correct.py    From NiMARE with MIT License 6 votes vote down vote up 
def _transform(self, result):
        p = result.maps['p']
        _, p_corr = mc.fdrcorrection(p, alpha=self.alpha, method=self.method,
                                     is_sorted=False)
        corr_maps = {'p': p_corr}
        self._generate_secondary_maps(result, corr_maps)
        return corr_maps
Example #4
Source File: safe.py    From safepy with GNU General Public License v3.0 5 votes vote down vote up 
def compute_pvalues_by_hypergeom(self, **kwargs):

        if kwargs:
            if 'verbose' in kwargs:
                self.verbose = kwargs['verbose']

            if self.verbose:
                print('Overwriting global settings:')
                for k in kwargs:
                    print('\t%s=%s' % (k, str(kwargs[k])))

        # Make sure that the settings are still valid
        self.validate_config()

        if self.verbose:
            print('Using the hypergeometric test to calculate enrichment...')

        # Nodes with not-NaN values in >= 1 attribute
        nodes_not_nan = np.any(~np.isnan(self.node2attribute), axis=1)

        # -- Number of nodes
        # n = self.graph.number_of_nodes()    # total
        n = np.sum(nodes_not_nan)    # with not-NaN values in >=1 attribute

        N = np.zeros([self.graph.number_of_nodes(), len(self.attributes)]) + n

        # -- Number of nodes annotated to each attribute
        N_in_group = np.tile(np.nansum(self.node2attribute, axis=0), (self.graph.number_of_nodes(), 1))

        # -- Number of nodes in each neighborhood
        # neighborhood_size = np.sum(self.neighborhoods, axis=0)[:, np.newaxis]    # total
        neighborhood_size = np.dot(self.neighborhoods,
                                   nodes_not_nan.astype(int))[:, np.newaxis] # with not-NaN values in >=1 attribute

        N_in_neighborhood = np.tile(neighborhood_size, (1, len(self.attributes)))

        # -- Number of nodes in each neighborhood and  annotated to each attribute
        N_in_neighborhood_in_group = np.dot(self.neighborhoods,
                                            np.where(~np.isnan(self.node2attribute), self.node2attribute, 0))

        self.pvalues_pos = hypergeom.sf(N_in_neighborhood_in_group - 1, N, N_in_group, N_in_neighborhood)

        # Correct for multiple testing
        if self.multiple_testing:

            if self.verbose:
                print('Running FDR-adjustment of p-values...')

            out = np.apply_along_axis(fdrcorrection, 1, self.pvalues_pos)
            self.pvalues_pos = out[:, 1, :]

        # Log-transform into neighborhood enrichment scores (NES)
        self.nes = -np.log10(self.pvalues_pos)