bsfit.nodes.models.bayes.CardinalBayes

class bsfit.nodes.models.bayes.CardinalBayes(initial_params: Dict[str, list], prior_shape: str, prior_mode: float, readout: str)

Bases: bsfit.nodes.models.bayes.StandardBayes

Cardinal Bayesian model

__init__(initial_params: Dict[str, list], prior_shape: str, prior_mode: float, readout: str)

Instantiate Cardinal Bayesian model

Parameters

• prior_shape (str) – prior shape

• "vonMisesPrior" (-) –

• prior_mode (float) – prior mode

• readout (str) – the posterior readout

• "map" (-) – maximum a posteriori

Exceptions:

MissingParameter: - “k_card” parameter is missing

Methods

__init__(initial_params, prior_shape, ...)	Instantiate Cardinal Bayesian model
fit(dataset)	fit the model
get_attributes()	get model attributes
predict(data, granularity)	generate the model's predictions
simulate(dataset, granularity, centering, ...)	simulate predictions

fit(dataset: pandas.core.frame.DataFrame)

fit the model

Parameters

• dataset (pd.DataFrame) –

  dataset with columns:

  'stim_mean': the stimulus feature (e.g., motion direction)
  'stim_std': the stimulus feature strength
  'prior_mode': mode of prior
  'prior_std': std of the prior
  'prior_shape': density function of the prior (e.g., "vonMisesPrior")
  'estimate': the estimate data
  

• init_p (dict) – initial parameters

Usage:

from bsfit.nodes.models.bayes import StandardBayes
initial_params = {
    "k_llh": [2.7, 10.7, 33],
    "k_prior": [2.7, 33],
    "p_rand": [0],
    "k_m": [2000],
}
model = StandardBayes(
    initial_params=initial_params,
    prior_shape="vonMisesPrior",
    prior_mode=225,
    readout="map")
model.fit(dataset)

Returns

the fitted model

Return type

(Model)

get_attributes()

get model attributes

Parameters

self (Model) – the model

Returns

list the model attributes

Return type

(list)

predict(data: tuple, granularity: str) → dict

generate the model’s predictions

Parameters

• data (tuple) – the data to predict, tuple of (stim_mean, stim_estimate)

• granularity (str) – (“trial” or “mean”) to predict single choices or their stats

Returns

the model predictions

Return type

(dict)

simulate(dataset: pandas.core.frame.DataFrame, granularity: str, centering: bool, **kwargs: dict)

simulate predictions

Parameters

• dataset (pd.DataFrame) –

  dataset:

  'stim_mean': the stimulus feature (e.g., motion direction)
  'stim_std': the stimulus feature strength
  'prior_mode': the stimulus feature density's mode
  'prior_std': the stimulus feature density's std
  'prior_shape': prior's density function (e.g., "vonMisesPrior")
  'estimate' (optional): the estimate data
  

• granularity (str) – (“trial” or “mean”)

• centering (bool) – center or not plot’s x-axis

Kwargs:

n_repeats (int): the number of repeats of conditions, when granularity=”trial”

Returns

results of the simulation

Return type

(dict)