def main(filename): plotname = filename[:-4] # remove ".csv" x, y, ids = util.process_csv_list(filename) heatmap, xedges, yedges = np.histogram2d(x, y, bins=(10, 4)) heatmap = heatmap[::-1] extent = [xedges[0], xedges[-1], yedges[0], yedges[-1]] plt.imshow(heatmap, extent=extent, interpolation='nearest') plt.colorbar() plt.savefig(plotname) print "Generated heatmap image %s.png" % plotname
def main(dirname): filenames = os.listdir(dirname) for filename in filenames: if not ".json" in filename: continue filename = filename[:-5] + ".csv" print "Processing %s... " % filename[:-4], x, y, ids = util.process_csv_list(os.path.join(dirname, filename)) new_x, new_y, removed = mahalanobis.remove_outliers(x, y) print "done" print "Removed %d outlying point(s)" % (len(x) - len(new_x)) print "Average error: %f, standard error: %f" % (rmse.calculate_rmse(new_x, new_y)) print
def make_arrays(dirname): data, predictions, boost_labels, maxent_labels = [], [], [], [] filenames = os.listdir(dirname) for filename in filenames: if not ".json" in filename: continue print "Collecting data from %s... " % filename[:-5], conditions = util.process_json(os.path.join(dirname, filename)) x, y, idx = util.process_csv_list( os.path.join(dirname, filename[:-5] + ".csv")) for i in range(len(x)): tempd, tempp = [0] * (util.num_conditions + 1), [0] * (util.num_conditions + 1) for j in conditions: tempd[j] = 1 tempp[j] = -1 data.append(tempd) predictions.append(tempp) error = rmse.get_single_rmse(x[i], y[i]) if error <= util.base_error / 2: boost_labels.append(1) maxent_labels.append(1) else: boost_labels.append(-1) maxent_labels.append(0) print "done" print data = np.array(data) predictions = np.array(predictions) boost_labels = np.array(boost_labels) maxent_labels = np.array(maxent_labels) return (data, predictions, boost_labels, maxent_labels)
def make_arrays(dirname): data, predictions, boost_labels, maxent_labels = [], [], [], [] filenames = os.listdir(dirname) for filename in filenames: if not ".json" in filename: continue print "Collecting data from %s... " % filename[:-5], conditions = util.process_json(os.path.join(dirname, filename)) x, y, idx = util.process_csv_list(os.path.join(dirname, filename[:-5] + ".csv")) for i in range(len(x)): tempd, tempp = [0] * (util.num_conditions + 1), [0] * (util.num_conditions + 1) for j in conditions: tempd[j] = 1 tempp[j] = -1 data.append(tempd) predictions.append(tempp) error = rmse.get_single_rmse(x[i], y[i]) if error <= util.base_error / 2: boost_labels.append(1) maxent_labels.append(1) else: boost_labels.append(-1) maxent_labels.append(0) print "done" print data = np.array(data) predictions = np.array(predictions) boost_labels = np.array(boost_labels) maxent_labels = np.array(maxent_labels) return (data, predictions, boost_labels, maxent_labels)
def main(filename): actual, data = util.process_csv_list(filename) print "average error: %f \nstandard error: %f" % (calculate_rmse( actual, data))
def main(filename): x, y = util.process_csv_list(filename) dm = calculate_mahalanobis(x, y) dm.sort() print dm
def main(filename): actual, data = util.process_csv_list(filename) print "average error: %f \nstandard error: %f" % (calculate_rmse(actual, data))