ecoli.library.data_predicates
Data Predicates
Defines several assertions about data that are useful for tests, e.g. checks for monotonicity, whether the data approximately follows a Poisson distribution, etc.
All functions expect a 1D numpy array as first parameter.
TODO: - implement faster numpy-based solution for tests of increasing/decreasing
- ecoli.library.data_predicates.approx_poisson(data, rate=None, significance=0.05, verbose=False)[source]
Test whether data appears to follow Poisson distribution, using Chi-sq goodness of fit. Does not do particularly well comparing poisson data of rate r_1 vs. poisson distribution of rate r_2. :param data: 1D array where index i corresponds the number of events observed in interval i. :param rate: rate (lambda) of the Poisson distribution against which to compare. If None, rate is estimated from the data. :param significance: for p > significance, fail to reject that the data is not Poisson-distributed. :param verbose: if True, prints estimated rate, and results (chi-sq, p-value) of the goodness-of-fit test.