我们从Python开源项目中，提取了以下50个代码示例，用于说明如何使用numpy.nansum()。
def makedists(pdata,binl): ##### This is called from within makeraindist. ##### Caclulate distributions pds=pdata.shape; nlat=pds[1]; nlon=pds[0]; nd=pds[2] bins=np.append(0,binl) n=np.empty((nlon,nlat,len(binl))) binno=np.empty(pdata.shape) for ilon in range(nlon): for ilat in range(nlat): # this is the histogram - we'll get frequency from this thisn,thisbin=np.histogram(pdata[ilon,ilat,:],bins) n[ilon,ilat,:]=thisn # these are the bin locations. we'll use these for the amount dist binno[ilon,ilat,:]=np.digitize(pdata[ilon,ilat,:],bins) #### Calculate the number of days with non-missing data, for normalization ndmat=np.tile(np.expand_dims(np.nansum(n,axis=2),axis=2),(1,1,len(bins)-1)) thisppdfmap=n/ndmat #### Iterate back over the bins and add up all the precip - this will be the rain amount distribution testpamtmap=np.empty(thisppdfmap.shape) for ibin in range(len(bins)-1): testpamtmap[:,:,ibin]=(pdata*(ibin==binno)).sum(axis=2) thispamtmap=testpamtmap/ndmat return thisppdfmap,thispamtmap
def plot_power_rose(wind_directions,power,num_wd_bins): """Plot a power rose. Kind of a hacked wind rose. Arguments: wind_directions -- a np array of wind directions filtered for icing power -- a np array of percent power production corresponding to wind_directions num_wd_bins -- the number of wind direction bins to include on the rose. """ dir_bins = np.array(np.linspace(0.0,360.0 - 360.0 / num_wd_bins,num_wd_bins)) #Find the total amount of power produced in each sector. dir_power = np.array([np.nansum(filter_obstacles(power,wind_directions,(wd + 180.0) % 360.0, 360 - 360/float(num_wd_bins))) for wd in dir_bins]) dir_power = np.round(dir_power * 100.0 / np.nansum(dir_power), decimals=0) #Normalize it and round to nearest int. proportional_wd = np.array([]) for i in range(len(dir_power)): for n in range(int(dir_power[i])): #Loop as many times as the percent of power produced in this sector. proportional_wd = np.append(proportional_wd,dir_bins[i]) #i.e., if 50% of power comes from the south, append 50 instances of 180.0 degrees. ones = np.ones(len(proportional_wd)) ax = new_axes() ax.bar(proportional_wd, ones,normed=False, opening=0.8, edgecolor='white', bins = [0.0,100.], cmap=cm.RdGy) set_legend(ax)
def ests_ll_quad(self, params): """ Calculate the loglikelihood given model parameters `params`. This method uses Gaussian quadrature, and thus returns an *approximate* integral. """ mu0, gamma0, err0 = np.split(params, 3) x = np.tile(self.z, (self.cfg.QCOUNT, 1, 1)) # (QCOUNTXnhospXnmeas) loc = mu0 + np.outer(QC1, gamma0) loc = np.tile(loc, (self.n, 1, 1)) loc = np.transpose(loc, (1, 0, 2)) scale = np.tile(err0, (self.cfg.QCOUNT, self.n, 1)) zs = lpdf_3d(x=x, loc=loc, scale=scale) w2 = np.tile(self.w, (self.cfg.QCOUNT, 1, 1)) wted = np.nansum(w2 * zs, axis=2).T # (nhosp X QCOUNT) qh = np.tile(QC1, (self.n, 1)) # (nhosp X QCOUNT) combined = wted + norm.logpdf(qh) # (nhosp X QCOUNT) return logsumexp(np.nan_to_num(combined), b=QC2, axis=1) # (nhosp)
def chi2(b, dataset, model1='phoebe1model', model2='phoebe2model'): ds = b.get_dataset(dataset) - b.get_dataset(dataset, method='*dep') if ds.method=='lc': depvar = 'fluxes' elif ds.method=='rv': depvar = 'rvs' else: raise NotImplementedError("chi2 doesn't support dataset method: '{}'".format(ds.method)) chi2 = 0.0 for comp in ds.components if len(ds.components) else [None]: if comp=='_default': continue # phoebe gives nans for RVs when a star is completely eclipsed, whereas # phoebe1 will give a value. So let's use nansum to just ignore those # regions of the RV curve print "***", depvar, dataset, model1, model2, comp chi2 += np.nansum((b.get_value(qualifier=depvar, dataset=dataset, model=model1, component=comp, context='model')\ -b.get_value(qualifier=depvar, dataset=dataset, model=model2, component=comp, context='model'))**2) return chi2
def weighted_average(weights, pep_abd, group_ix): ''' Calculate weighted geometric means for sample groups Inputs: weights: weights of peptides after filtering by loading threshold pep_abd: peptide abundances after filtering by loading threshold group_ix: array indexes of sample groups ''' global nGroups abd_w = pep_abd * weights[..., None] one_w = abd_w / abd_w * weights[..., None] a_sums = np.nansum(abd_w, axis=0) w_sums = np.nansum(one_w, axis=0) expr = np.empty(nGroups) for i in range(expr.shape[0]): expr[i] = a_sums[group_ix[i]].sum() / w_sums[group_ix[i]].sum() return expr
def pwdist_canberra(self, seq1idx, seq2idx): """Compute the Canberra distance between two vectors. References: 1. http://scipy.org/ Notes: When `u[i]` and `v[i]` are 0 for given i, then the fraction 0/0 = 0 is used in the calculation. """ u = self[seq1idx] v = self[seq2idx] olderr = np.seterr(invalid='ignore') try: d = np.nansum(abs(u - v) / (abs(u) + abs(v))) finally: np.seterr(**olderr) return d
def normalize(self, to=1.0): """ This function ... :param to: :return: """ # Calculate the sum of all the pixels sum = np.nansum(self) # Calculate the conversion factor factor = to / sum # Multiply the frame with the conversion factor self.__imul__(factor) # -----------------------------------------------------------------
def calculate_optimizer_time(trials): optimizer_time = [] time_idx = 0 optimizer_time.append(trials.cv_starttime[0] - trials.starttime[time_idx]) for i in range(len(trials.cv_starttime[1:])): if trials.cv_starttime[i + 1] > trials.endtime[time_idx]: optimizer_time.append(trials.endtime[time_idx] - trials.cv_endtime[i]) time_idx += 1 optimizer_time.append(trials.cv_starttime[i + 1] - trials.starttime[time_idx]) else: optimizer_time.append(trials.cv_starttime[i + 1] - trials.cv_endtime[i]) optimizer_time.append(trials.endtime[time_idx] - trials.cv_endtime[-1]) trials.optimizer_time = optimizer_time # We need to import numpy again import numpy as np return np.nansum(optimizer_time)
def lnlike(self, pars): # Pull theta out of pars theta = pars[:self.Nbins] # Generate the inner summation gamma = np.ones_like(self.bin_idx) * np.nan good = (self.bin_idx < self.Nbins) & (self.bin_idx >= 0) # nans in q get put in nonexistent bins gamma[good] = self.Nobs * self.censoring_fcn(self.mcmc_samples[good]) * theta[self.bin_idx[good]] summation = np.nanmean(gamma, axis=1) # Calculate the integral I = self._integral_fcn(theta) # Generate the log-likelihood ll = -I + np.nansum(np.log(summation)) return ll
def test_sum_inf(self): import pandas.core.nanops as nanops s = Series(np.random.randn(10)) s2 = s.copy() s[5:8] = np.inf s2[5:8] = np.nan self.assertTrue(np.isinf(s.sum())) arr = np.random.randn(100, 100).astype('f4') arr[:, 2] = np.inf with cf.option_context("mode.use_inf_as_null", True): assert_almost_equal(s.sum(), s2.sum()) res = nanops.nansum(arr, axis=1) self.assertTrue(np.isinf(res).all())
def r(self): """ Pearson correlation of the fitted Variogram :return: """ # get the experimental and theoretical variogram and cacluate means experimental, model = self.__model_deviations() mx = np.nanmean(experimental) my = np.nanmean(model) # claculate the single pearson correlation terms term1 = np.nansum(np.fromiter(map(lambda x, y: (x-mx) * (y-my), experimental, model), np.float)) t2x = np.nansum(np.fromiter(map(lambda x: (x-mx)**2, experimental), np.float)) t2y = np.nansum(np.fromiter(map(lambda y: (y-my)**2, model), np.float)) return term1 / (np.sqrt(t2x * t2y))
def entropy(v, axis=0): """ Optimized implementation of entropy. This version is faster than that in scipy.stats.distributions, particularly over long vectors. """ v = numpy.array(v, dtype='float') s = numpy.sum(v, axis=axis) with numpy.errstate(divide='ignore', invalid='ignore'): rhs = numpy.nansum(v * numpy.log(v), axis=axis) / s r = numpy.log(s) - rhs # Where dealing with binarized events, it is possible that an event always # occurs and thus has 0 information. In this case, the negative class # will have frequency 0, resulting in log(0) being computed as nan. # We replace these nans with 0 nan_index = numpy.isnan(rhs) if nan_index.any(): r[nan_index] = 0 return r
def __call__(self, y_true_proba, y_proba): """ See Murphy (1973) A vector partition of the probability score """ np.seterr(divide="ignore") pos_obs_freq = np.histogram( y_proba[y_true_proba == 1], bins=self.bins)[0] fore_freq = np.histogram(y_proba, bins=self.bins)[0] climo = y_true_proba.mean() unc = climo * (1 - climo) pos_obs_rel_freq = np.zeros(pos_obs_freq.size) for p in range(pos_obs_rel_freq.size): if fore_freq[p] > 0: pos_obs_rel_freq[p] = pos_obs_freq[p] / fore_freq[p] else: pos_obs_rel_freq[p] = np.nan score = np.nansum(fore_freq * (pos_obs_rel_freq - climo) ** 2) score /= float(y_proba.size) return score / unc
def cluster_f_measure(ytrue, pred): # higher is better assert len(ytrue) == len(pred), 'inputs length must be equal.' label2ix = {label: i for i, label in enumerate(np.unique(ytrue))} _ytrue = np.array([label2ix[v] for v in ytrue]) nSize = len(_ytrue) nClassTrue = len(np.unique(ytrue)) nClassPred = len(np.unique(pred)) f = np.zeros((nClassTrue, nClassPred)).astype(dtype=np.float64) for i in xrange(nClassTrue): freq_i = len(_ytrue[_ytrue == i]) for j in xrange(nClassPred): freq_j = len(pred[pred == j]) freq_i_j = float(len(filter(lambda x: x == j, pred[_ytrue == i]))) precision = freq_i_j / freq_j if freq_j != 0 else 0 recall = freq_i_j / freq_i if freq_i != 0 else 0 if precision == 0 or recall == 0: f[i, j] = 0. else: f[i, j] = 2. * (precision * recall) / (precision + recall) return np.nansum([f[i][j] * len(_ytrue[_ytrue == i]) for i in xrange(nClassTrue) for j in xrange(nClassPred)]) / nSize
def ponderateByConcentration(): print 'Loading feature concentration..........' sdFile = open('varStandarDevs.txt','rb') standevs=pickle.load(sdFile) sdFile.close() totDevs={} for feature in standevs: totDevs[feature]=sum([abs(standevs[feature][si]) for si in range(len(standevs[feature]))])/len(standevs[feature]) localF=['turningAngle','turningAngleDifference','Coord','LP'] globalF=['accAngle','coG','relStrokeLength','liS','quadraticError'] totalF=['turningAngle','turningAngleDifference','Coord','LP','Style','accAngle','coG','relStrokeLength','liS','quadraticError'] print 'Ponderating features..........' weights={} norm=np.nansum([1/float(math.sqrt(totDevs[feature])) for feature in totalF]) for feature in totalF: weights[feature]=(1/float(math.sqrt(totDevs[feature])))/float(norm) print 'Features weighted as' print weights return weights
def nandot(a, b): # TODO: speed up, avoid copying data "A numpy.dot() replacement which treats (0*-Inf)==0 and works around BLAS NaN bugs in matrices." # important note: a contains zeros and b contains inf/-inf/nan, not the other way around # workaround for zero*-inf=nan in dot product (must be 0 according to 0^0=1 with probabilities) # 1) calculate dot product # 2) select nan entries # 3) re-calculate matrix entries where 0*inf = 0 using np.nansum() tmp = np.dot(a, b) indices = np.where(np.isnan(tmp)) ri, ci = indices with np.errstate(invalid='ignore'): values = np.nansum(a[ri, :] * b[:, ci].T, axis=1) values[np.isnan(values)] = 0.0 tmp[indices] = values return tmp
def argnanmedoid(x, axis=1): """ Return the indices of the medoid :param x: input array :param axis: axis to medoid along :return: indices of the medoid """ if axis == 0: x = x.T invalid = anynan(x, axis=0) band, time = x.shape diff = x.reshape(band, time, 1) - x.reshape(band, 1, time) dist = np.sqrt(np.sum(diff * diff, axis=0)) # dist = np.linalg.norm(diff, axis=0) is slower somehow... dist_sum = nansum(dist, axis=0) dist_sum[invalid] = np.inf i = np.argmin(dist_sum) return i
def medoid_indices(arr, invalid=None): """ The indices of the medoid. :arg arr: input array :arg invalid: mask for invalid data containing NaNs """ # vectorized version of `argnanmedoid` bands, times, ys, xs = arr.shape diff = (arr.reshape(bands, times, 1, ys, xs) - arr.reshape(bands, 1, times, ys, xs)) dist = np.linalg.norm(diff, axis=0) dist_sum = nansum(dist, axis=0) if invalid is None: # compute it in case it's not already available invalid = anynan(arr, axis=0) dist_sum[invalid] = np.inf return np.argmin(dist_sum, axis=0)
def frame_to_series(self, field, frame, columns=None): """ Convert a frame with a DatetimeIndex and sid columns into a series with a sid index, using the aggregator defined by the given field. """ if isinstance(frame, pd.DataFrame): columns = frame.columns frame = frame.values if not len(frame): return pd.Series( data=(0 if field == 'volume' else np.nan), index=columns, ).values if field in ['price', 'close']: # shortcircuit for full last row vals = frame[-1] if np.all(~np.isnan(vals)): return vals return ffill(frame)[-1] elif field == 'open': return bfill(frame)[0] elif field == 'volume': return np.nansum(frame, axis=0) elif field == 'high': return np.nanmax(frame, axis=0) elif field == 'low': return np.nanmin(frame, axis=0) else: raise ValueError("Unknown field {}".format(field))
def getJointNumFramesVisible(self, jointID): """ Get number of frames in which joint is visible :param jointID: joint ID :return: number of frames """ return numpy.nansum(self.gt[:, jointID, :]) / self.gt.shape[2] # 3D
def est_pmf_from_mpps(self, other, samples, eps=1e-10): """Estimate probability mass function from MPPovmList samples :param MPPovmList other: An :class:`MPPovmList` instance :param samples: Iterable of samples (e.g. from :func:`MPPovmList.samples()`) :returns: `(p_est, n_samples_used)`, both are shape `self.nsoutdims` ndarrays. `p_est` provides estimated probabilities and `n_samples_used` provides the effective number of samples used for each probability. """ assert len(other.mpps) == len(samples) pmf_ests = np.zeros((len(other.mpps),) + self.nsoutdims, float) n_samples = np.zeros(len(other.mpps), int) for pos, other_mpp, other_samples in zip(it.count(), other.mpps, samples): pmf_ests[pos, ...], n_samples[pos] = self.est_pmf_from( other_mpp, other_samples, eps) n_out = np.prod(self.nsoutdims) pmf_ests = pmf_ests.reshape((len(other.mpps), n_out)) given = ~np.isnan(pmf_ests) n_samples_used = (given * n_samples[:, None]).sum(0) # Weighted average over available estimates according to the # number of samples underlying each estimate. Probabilities # without any estimates produce 0.0 / 0 = nan in `pmf_est`. pmf_est = np.nansum(pmf_ests * n_samples[:, None], 0) / n_samples_used return (pmf_est.reshape(self.nsoutdims), n_samples_used.reshape(self.nsoutdims))
def test_nansum_with_boolean(self): # gh-2978 a = np.zeros(2, dtype=np.bool) try: np.nansum(a) except: raise AssertionError()
def test_nansum(self): tgt = np.sum(self.mat) for mat in self.integer_arrays(): assert_equal(np.nansum(mat), tgt)
def test_allnans(self): # Check for FutureWarning with warnings.catch_warnings(record=True) as w: warnings.simplefilter('always') res = np.nansum([np.nan]*3, axis=None) assert_(res == 0, 'result is not 0') assert_(len(w) == 0, 'warning raised') # Check scalar res = np.nansum(np.nan) assert_(res == 0, 'result is not 0') assert_(len(w) == 0, 'warning raised') # Check there is no warning for not all-nan np.nansum([0]*3, axis=None) assert_(len(w) == 0, 'unwanted warning raised')
def test_empty(self): for f, tgt_value in zip([np.nansum, np.nanprod], [0, 1]): mat = np.zeros((0, 3)) tgt = [tgt_value]*3 res = f(mat, axis=0) assert_equal(res, tgt) tgt = [] res = f(mat, axis=1) assert_equal(res, tgt) tgt = tgt_value res = f(mat, axis=None) assert_equal(res, tgt)
def compact_logit(x, eps=.00001): import warnings with warnings.catch_warnings(): warnings.filterwarnings("ignore", message="divide by zero encountered in true_divide") warnings.filterwarnings("ignore", message="divide by zero encountered in log") warnings.filterwarnings("ignore", message="invalid value encountered in multiply") return np.nansum(((x<=eps)*x, (x>=(1-eps))*x, ((x>eps)&(x<(1-eps)))*((1-2*eps)*(np.log(x/(1-x)))/(2*np.log((1-eps)/eps))+.5)),axis=0)
def _get_weights(self, data, kpi, variant): if kpi not in self.reference_kpis: return 1.0 reference_kpi = self.reference_kpis[kpi] x = self.get_kpi_by_name_and_variant(data, reference_kpi, variant) zeros_and_nans = sum(x == 0) + np.isnan(x).sum() non_zeros = len(x) - zeros_and_nans return non_zeros/np.nansum(x) * x
def KL(a, b): """Calculate the Kullback Leibler divergence between a and b """ D_KL = np.nansum(np.multiply(a, np.log(np.divide(a, b+np.spacing(1)))), axis=1) return D_KL
def calc_information(probTgivenXs, PYgivenTs, PXs, PYs): """Calculate the MI - I(X;T) and I(Y;T)""" PTs = np.nansum(probTgivenXs*PXs, axis=1) Ht = np.nansum(-np.dot(PTs, np.log2(PTs))) Htx = - np.nansum((np.dot(np.multiply(probTgivenXs, np.log2(probTgivenXs)), PXs))) Hyt = - np.nansum(np.dot(PYgivenTs*np.log2(PYgivenTs+np.spacing(1)), PTs)) Hy = np.nansum(-PYs * np.log2(PYs+np.spacing(1))) IYT = Hy - Hyt ITX = Ht - Htx return ITX, IYT
def calc_information_1(probTgivenXs, PYgivenTs, PXs, PYs, PTs): """Calculate the MI - I(X;T) and I(Y;T)""" #PTs = np.nansum(probTgivenXs*PXs, axis=1) Ht = np.nansum(-np.dot(PTs, np.log2(PTs+np.spacing(1)))) Htx = - np.nansum((np.dot(np.multiply(probTgivenXs, np.log2(probTgivenXs+np.spacing(1))), PXs))) Hyt = - np.nansum(np.dot(PYgivenTs*np.log2(PYgivenTs+np.spacing(1)), PTs)) Hy = np.nansum(-PYs * np.log2(PYs+np.spacing(1))) IYT = Hy - Hyt ITX = Ht - Htx return ITX, IYT
def calc_information(probTgivenXs, PYgivenTs, PXs, PYs, PTs): """Calculate the MI - I(X;T) and I(Y;T)""" #PTs = np.nansum(probTgivenXs*PXs, axis=1) t_indeces = np.nonzero(PTs) Ht = np.nansum(-np.dot(PTs, np.log2(PTs+np.spacing(1)))) Htx = - np.nansum((np.dot(np.multiply(probTgivenXs, np.log2(probTgivenXs)), PXs))) Hyt = - np.nansum(np.dot(PYgivenTs*np.log2(PYgivenTs+np.spacing(1)), PTs)) Hy = np.nansum(-PYs * np.log2(PYs+np.spacing(1))) IYT = Hy - Hyt ITX = Ht - Htx return ITX, IYT
def t_calc_information(p_x_given_t, PYgivenTs, PXs, PYs): """Calculate the MI - I(X;T) and I(Y;T)""" Hx = np.nansum(-np.dot(PXs, np.log2(PXs))) Hxt = - np.nansum((np.dot(np.multiply(p_x_given_t, np.log2(p_x_given_t)), PXs))) Hyt = - np.nansum(np.dot(PYgivenTs*np.log2(PYgivenTs+np.spacing(1)), PTs)) Hy = np.nansum(-PYs * np.log2(PYs+np.spacing(1))) IYT = Hy - Hyt ITX = Hx - Hxt return ITX, IYT
def _fit_cdd_only(df, weighted=False): bps = [i[4:] for i in df.columns if i[:3] == 'CDD'] best_bp, best_rsquared, best_mod, best_res = None, -9e9, None, None best_formula, cdd_qualified = None, False try: # TODO: fix big try block anti-pattern for bp in bps: candidate_cdd_formula = 'upd ~ CDD_' + bp if (np.nansum(df['CDD_' + bp] > 0) < 10) or \ (np.nansum(df['CDD_' + bp]) < 20): continue if weighted: candidate_cdd_mod = smf.wls(formula=candidate_cdd_formula, data=df, weights=df['ndays']) else: candidate_cdd_mod = smf.ols(formula=candidate_cdd_formula, data=df) candidate_cdd_res = candidate_cdd_mod.fit() candidate_cdd_rsquared = candidate_cdd_res.rsquared_adj if (candidate_cdd_rsquared > best_rsquared and candidate_cdd_res.params['Intercept'] >= 0 and candidate_cdd_res.params['CDD_' + bp] >= 0 and candidate_cdd_res.pvalues['CDD_' + bp] < 0.1): best_bp, best_rsquared = int(bp), candidate_cdd_rsquared best_mod, best_res = candidate_cdd_mod, candidate_cdd_res cdd_qualified = True best_formula = 'upd ~ CDD_' + bp except: # TODO: catch specific error best_rsquared, cdd_qualified = 0, False best_formula, best_mod, best_res = None, None, None best_bp = None return best_formula, best_mod, best_res, best_rsquared, cdd_qualified, best_bp
def _fit_hdd_only(df, weighted=False): bps = [i[4:] for i in df.columns if i[:3] == 'HDD'] best_bp, best_rsquared, best_mod, best_res = None, -9e9, None, None best_formula, hdd_qualified = None, False try: # TODO: fix big try block anti-pattern for bp in bps: candidate_hdd_formula = 'upd ~ HDD_' + bp if (np.nansum(df['HDD_' + bp] > 0) < 10) or \ (np.nansum(df['HDD_' + bp]) < 20): continue if weighted: candidate_hdd_mod = smf.wls(formula=candidate_hdd_formula, data=df, weights=df['ndays']) else: candidate_hdd_mod = smf.ols(formula=candidate_hdd_formula, data=df) candidate_hdd_res = candidate_hdd_mod.fit() candidate_hdd_rsquared = candidate_hdd_res.rsquared_adj if (candidate_hdd_rsquared > best_rsquared and candidate_hdd_res.params['Intercept'] >= 0 and candidate_hdd_res.params['HDD_' + bp] >= 0 and candidate_hdd_res.pvalues['HDD_' + bp] < 0.1): best_bp, best_rsquared = int(bp), candidate_hdd_rsquared best_mod, best_res = candidate_hdd_mod, candidate_hdd_res hdd_qualified = True best_formula = 'upd ~ HDD_' + bp except: # TODO: catch specific error best_rsquared, hdd_qualified = 0, False best_formula, best_mod, best_res = None, None, None best_bp = None return best_formula, best_mod, best_res, best_rsquared, hdd_qualified, best_bp
def calc_gross(self): return np.nansum(self.input_data.energy)
def get_relevance_scores(matched_predictions, positive_feedback, not_rated_penalty): users_num = matched_predictions.shape[0] reldata = get_relevance_data(matched_predictions, positive_feedback, not_rated_penalty) true_pos, false_pos = reldata.tp, reldata.fp true_neg, false_neg = reldata.tn, reldata.fn with np.errstate(invalid='ignore'): # true positive rate precision = true_pos / (true_pos + false_pos) # sensitivity recall = true_pos / (true_pos + false_neg) # false positive rate fallout = false_pos / (false_pos + true_neg) # true negative rate specifity = true_neg / (false_pos + true_neg) # false negative rate miss_rate = false_neg / (false_neg + true_pos) #average over all users precision = unmask(np.nansum(precision) / users_num) recall = unmask(np.nansum(recall) / users_num) fallout = unmask(np.nansum(fallout) / users_num) specifity = unmask(np.nansum(specifity) / users_num) miss_rate = unmask(np.nansum(miss_rate) / users_num) scores = namedtuple('Relevance', ['precision', 'recall', 'fallout', 'specifity', 'miss_rate']) scores = scores._make([precision, recall, fallout, specifity, miss_rate]) return scores
def get_ranking_scores(matched_predictions, feedback_data, switch_positive, alternative=True): users_num, topk, holdout = matched_predictions.shape ideal_scores_idx = np.argsort(feedback_data, axis=1)[:, ::-1] #returns column index only ideal_scores_idx = np.ravel_multi_index((np.arange(feedback_data.shape[0])[:, None], ideal_scores_idx), dims=feedback_data.shape) where = np.ma.where if np.ma.is_masked(feedback_data) else np.where is_positive = feedback_data >= switch_positive positive_feedback = where(is_positive, feedback_data, 0) negative_feedback = where(~is_positive, -feedback_data, 0) relevance_scores_pos = (matched_predictions * positive_feedback[:, None, :]).sum(axis=2) relevance_scores_neg = (matched_predictions * negative_feedback[:, None, :]).sum(axis=2) ideal_scores_pos = positive_feedback.ravel()[ideal_scores_idx] ideal_scores_neg = negative_feedback.ravel()[ideal_scores_idx] discount_num = max(holdout, topk) if alternative: discount = np.log2(np.arange(2, discount_num+2)) relevance_scores_pos = 2**relevance_scores_pos - 1 relevance_scores_neg = 2**relevance_scores_neg - 1 ideal_scores_pos = 2**ideal_scores_pos - 1 ideal_scores_neg = 2**ideal_scores_neg - 1 else: discount = np.hstack([1, np.log(np.arange(2, discount_num+1))]) dcg = (relevance_scores_pos / discount[:topk]).sum(axis=1) dcl = (relevance_scores_neg / -discount[:topk]).sum(axis=1) idcg = (ideal_scores_pos / discount[:holdout]).sum(axis=1) idcl = (ideal_scores_neg / -discount[:holdout]).sum(axis=1) with np.errstate(invalid='ignore'): ndcg = unmask(np.nansum(dcg / idcg) / users_num) ndcl = unmask(np.nansum(dcl / idcl) / users_num) ranking_score = namedtuple('Ranking', ['nDCG', 'nDCL'])._make([ndcg, ndcl]) return ranking_score
def vwap(df): """ Volume-weighted average price (VWAP) is a ratio generally used by institutional investors and mutual funds to make buys and sells so as not to disturb the market prices with large orders. It is the average share price of a stock weighted against its trading volume within a particular time frame, generally one day. Read more: Volume Weighted Average Price - VWAP https://www.investopedia.com/terms/v/vwap.asp#ixzz4xt922daE Parameters ---------- df: pd.DataFrame Returns ------- """ if 'close' not in df.columns or 'volume' not in df.columns: raise ValueError('price data must include `volume` and `close`') vol_sum = np.nansum(df['volume'].values) try: ret = np.nansum(df['close'].values * df['volume'].values) / vol_sum except ZeroDivisionError: ret = np.nan return ret
def _calculate(self, X, y, categorical, metafeatures, helpers): res = np.nansum(helpers.get_value("NumSymbols")) return res if np.isfinite(res) else 0 ################################################################################ # Statistical meta features # Only use third and fourth statistical moment because it is common to # standardize for the other two # see Engels & Theusinger, 1998 - Using a Data Metric for Preprocessing Advice for Data Mining Applications.
def trajectory_score_array(posterior, slope=None, intercept=None, w=None, weights=None, normalize=False): """Docstring goes here This is the score that Davidson et al. maximizes, in order to get a linear trajectory, but here we kind of assume that that we have the trajectory already, and then just score it. w is the number of bin rows to include in score, in each direction. That is, w=0 is only the modes, and w=1 is a band of width=3, namely the modes, and 1 bin above, and 1 bin below the mode. The score is NOT averaged!""" rows, cols = posterior.shape if w is None: w = 0 if not float(w).is_integer: raise ValueError("w has to be an integer!") if slope is None or intercept is None: slope, intercept, _ = linregress_array(posterior=posterior) x = np.arange(cols) line_y = np.round((slope*x + intercept)) # in position bin #s # idea: cycle each column so that the top w rows are the band surrounding the regression line if np.isnan(slope): # this will happen if we have 0 or only 1 decoded bins return np.nan else: temp = column_cycle_array(posterior, -line_y+w) if normalize: num_non_nan_bins = round(np.nansum(posterior)) else: num_non_nan_bins = 1 return np.nansum(temp[:2*w+1,:])/num_non_nan_bins
def test_nsum(x): assume(np.max(x[np.isfinite(x)]) < 1e4) assume(np.min(x[np.isfinite(x)]) > -1e4) aae(nsum(x), np.nansum(x))
def test_nsum_row(x): assume(np.max(x[np.isfinite(x)]) < 1e4) assume(np.min(x[np.isfinite(x)]) > -1e4) aae(nsum_row(x), np.nansum(x, axis=1))
def test_preds_ll(alpha, mu, gamma, err, num, w): current_impl = Lvm.preds_ll(alpha, mu, gamma, err, num, w) simple_impl = np.nansum(w * norm.logpdf(num, mu+gamma*alpha, err)) simple_impl += np.sum(norm.logpdf(alpha)) assert_approx_equal(current_impl, simple_impl)
def ests_obj(self, params): """The objective function to minimize for the model parameters.""" # return -nsum(self.ests_ll(params)) return -np.nansum(self.ests_ll(params))
def nsum_row(a): return nansum(a, axis=1)
def test_basic_stats(x): s = SummaryStats() s.update(x) assert s.count() == np.count_nonzero(~np.isnan(x)) np.testing.assert_allclose(s.sum(), np.nansum(x), rtol=RTOL, atol=ATOL) np.testing.assert_equal(s.min(), np.nanmin(x) if len(x) else np.nan) np.testing.assert_equal(s.max(), np.nanmax(x) if len(x) else np.nan) np.testing.assert_allclose(s.mean(), np.nanmean(x) if len(x) else np.nan, rtol=RTOL, atol=ATOL) np.testing.assert_allclose(s.var(), np.nanvar(x) if len(x) else np.nan, rtol=RTOL, atol=ATOL) np.testing.assert_allclose(s.std(), np.nanstd(x) if len(x) else np.nan, rtol=RTOL, atol=ATOL)
def log_likelihood(y, yhat): '''Helper function to compute the log likelihood.''' eps = np.spacing(1) return np.nansum(y * np.log(eps + yhat) - yhat)
