Colab Instructions¶

If you would like to run this tutorial on Google colab, please click this link.

Once you are in the colab, follow the installation instructions below and then restart your runtime.

Just uncomment the code in the next code cell and run it!

NOTE:

You may want to switch your runtime to have a GPU or TPU. To do so, go to Runtime > Change runtime type and select the desired hardware accelerator.

Note that if you switch your runtime you have to follow the installation instructions again.

In [1]:

# If running this on Colab, please uncomment the next line
# !pip install hssm

# If running this on Colab, please uncomment the next line # !pip install hssm

Using lapse probabilities and distributions in HSSM¶

Since v0.1.2, HSSM has added the ability to model outliers in the distribution with lapse probabilities and distributions. Let's see how it works.

In [2]:

import bambi as bmb
import matplotlib.pyplot as plt

import hssm

hssm.set_floatX("float32")

import warnings

warnings.filterwarnings("ignore")

import bambi as bmb import matplotlib.pyplot as plt import hssm hssm.set_floatX("float32") import warnings warnings.filterwarnings("ignore")

WARNING (pytensor.tensor.blas): Using NumPy C-API based implementation for BLAS functions.

Setting PyTensor floatX type to float32.

Setting "jax_enable_x64" to False. If this is not intended, please set `jax` to False.

Lapse probabilities and distributions are enabled by default.¶

From v0.1.2 on, lapse probabilities and distributions are enabled by default. If left unspecified, the probability for outliers will be a fixed value of 0.05 and the distribution will be specified as Uniform(0, 10).

In [3]:

# Simulate some data
ddm_data = hssm.simulate_data(
    model="ddm", theta=dict(v=0.5, a=1.5, z=0.5, t=0.1), size=1000
)
ddm_data.head()

# Simulate some data ddm_data = hssm.simulate_data( model="ddm", theta=dict(v=0.5, a=1.5, z=0.5, t=0.1), size=1000 ) ddm_data.head()

Out[3]:

	rt	response
0	2.447235	-1.0
1	3.504328	1.0
2	0.507807	-1.0
3	1.949458	1.0
4	0.753925	1.0

In [4]:

# Build the simplest model specifying only data
# Note the model output

ddm_model_default = hssm.HSSM(data=ddm_data)
ddm_model_default

# Build the simplest model specifying only data # Note the model output ddm_model_default = hssm.HSSM(data=ddm_data) ddm_model_default

Model initialized successfully.

Out[4]:

Hierarchical Sequential Sampling Model
Model: ddm

Response variable: rt,response
Likelihood: analytical
Observations: 1000

Parameters:

v:
    Prior: Normal(mu: 0.0, sigma: 2.0)
    Explicit bounds: (-inf, inf)

a:
    Prior: HalfNormal(sigma: 2.0)
    Explicit bounds: (0.0, inf)

z:
    Prior: Uniform(lower: 0.0, upper: 1.0)
    Explicit bounds: (0.0, 1.0)

t:
    Prior: HalfNormal(sigma: 2.0)
    Explicit bounds: (0.0, inf)


Lapse probability: 0.05
Lapse distribution: Uniform(lower: 0.0, upper: 20.0)

Note

Note in the above output that lapse probability and lapse distributions are at default values.

Specifying lapse probability and distribution¶

It is easy to change the lapse probability and distribution. HSSM has added two arguments, p_outlier and lapse to allow the lapse probability and distribution to be specified.

The optional p_outlier accepts a float, a dict, or a bmb.Prior object. When p_outlier is specified as a single float value, it will be considered "fixed". You can also specify a prior distribution for p_outlier through a dict or a bmb.Prior object, the same way as you would when specifying priors for any other parameter. That way, the lapse probability will be considered another parameter and will be estimated during MCMC sampling.

Likewise, the lapse argument accepts a dict or a bmb.Prior object to specify a fixed lapse distribution. This distribution will be considered as the one that outliers are generated from.

In [5]:

ddm_model_lapse = hssm.HSSM(
    data=ddm_data,
    p_outlier={"name": "Uniform", "lower": 0.0001, "upper": 0.5},
    lapse=bmb.Prior("Uniform", lower=0.0, upper=20.0),
)

ddm_model_lapse

ddm_model_lapse = hssm.HSSM( data=ddm_data, p_outlier={"name": "Uniform", "lower": 0.0001, "upper": 0.5}, lapse=bmb.Prior("Uniform", lower=0.0, upper=20.0), ) ddm_model_lapse

Model initialized successfully.

Out[5]:

Hierarchical Sequential Sampling Model
Model: ddm

Response variable: rt,response
Likelihood: analytical
Observations: 1000

Parameters:

v:
    Prior: Normal(mu: 0.0, sigma: 2.0)
    Explicit bounds: (-inf, inf)

a:
    Prior: HalfNormal(sigma: 2.0)
    Explicit bounds: (0.0, inf)

z:
    Prior: Uniform(lower: 0.0, upper: 1.0)
    Explicit bounds: (0.0, 1.0)

t:
    Prior: HalfNormal(sigma: 2.0)
    Explicit bounds: (0.0, inf)


Lapse probability: Uniform(lower: 9.999999747378752e-05, upper: 0.5)
Lapse distribution: Uniform(lower: 0.0, upper: 20.0)

In [6]:

lapse_trace = ddm_model_lapse.sample()
lapse_trace

lapse_trace = ddm_model_lapse.sample() lapse_trace

Using default initvals. 

Auto-assigning NUTS sampler...

Initializing NUTS using adapt_diag...

Multiprocess sampling (2 chains in 2 jobs)

NUTS: [t, z, p_outlier, a, v]

Sampling 2 chains for 1_000 tune and 1_000 draw iterations (2_000 + 2_000 draws total) took 30 seconds.

We recommend running at least 4 chains for robust computation of convergence diagnostics

  0%|          | 0/2000 [00:00<?, ?it/s]

  7%|▋         | 146/2000 [00:00<00:01, 1451.61it/s]

 15%|█▍        | 292/2000 [00:00<00:01, 1441.10it/s]

 22%|██▏       | 441/2000 [00:00<00:01, 1460.23it/s]

 29%|██▉       | 588/2000 [00:00<00:00, 1454.80it/s]

 37%|███▋      | 734/2000 [00:00<00:00, 1425.25it/s]

 44%|████▍     | 882/2000 [00:00<00:00, 1441.33it/s]

 52%|█████▏    | 1031/2000 [00:00<00:00, 1454.81it/s]

 59%|█████▉    | 1177/2000 [00:00<00:00, 1451.73it/s]

 66%|██████▌   | 1323/2000 [00:00<00:00, 1449.66it/s]

 73%|███████▎  | 1468/2000 [00:01<00:00, 1425.26it/s]

 81%|████████  | 1616/2000 [00:01<00:00, 1440.88it/s]

 88%|████████▊ | 1765/2000 [00:01<00:00, 1453.38it/s]

 96%|█████████▌| 1914/2000 [00:01<00:00, 1462.70it/s]

100%|██████████| 2000/2000 [00:01<00:00, 1448.12it/s]

Out[6]:

arviz.InferenceData

• posterior

  <xarray.Dataset> Size: 48kB
  Dimensions:    (chain: 2, draw: 1000)
  Coordinates:
    * chain      (chain) int64 16B 0 1
    * draw       (draw) int64 8kB 0 1 2 3 4 5 6 7 ... 993 994 995 996 997 998 999
  Data variables:
      z          (chain, draw) float32 8kB 0.5152 0.5146 0.492 ... 0.4937 0.5091
      v          (chain, draw) float32 8kB 0.4381 0.4907 0.5305 ... 0.5345 0.4804
      t          (chain, draw) float32 8kB 0.08865 0.112 ... 0.1044 0.09832
      p_outlier  (chain, draw) float32 8kB 0.00135 0.005336 ... 0.007374 0.008597
      a          (chain, draw) float32 8kB 1.573 1.524 1.579 ... 1.554 1.546 1.563
  Attributes:
      created_at:                  2024-09-18T23:33:36.552645+00:00
      arviz_version:               0.19.0
      inference_library:           pymc
      inference_library_version:   5.16.2
      sampling_time:               30.324628829956055
      tuning_steps:                1000
      modeling_interface:          bambi
      modeling_interface_version:  0.14.0

  xarray.Dataset

  • Dimensions:
    • chain: 2
    • draw: 1000
  • Coordinates: (2)
    • chain
      (chain)
      int64
      0 1

      array([0, 1])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

  • Data variables: (5)
    • z
      (chain, draw)
      float32
      0.5152 0.5146 ... 0.4937 0.5091

      array([[0.5151665 , 0.5145898 , 0.4920201 , ..., 0.492379  , 0.5190937 ,
              0.5037963 ],
             [0.49066404, 0.5100751 , 0.49091738, ..., 0.50943816, 0.4936943 ,
              0.5090597 ]], dtype=float32)

    • v
      (chain, draw)
      float32
      0.4381 0.4907 ... 0.5345 0.4804

      array([[0.43808955, 0.49065036, 0.5305067 , ..., 0.5167036 , 0.48403665,
              0.48575816],
             [0.5361283 , 0.5029262 , 0.4814368 , ..., 0.4533974 , 0.534528  ,
              0.48043993]], dtype=float32)

    • t
      (chain, draw)
      float32
      0.08865 0.112 ... 0.1044 0.09832

      array([[0.08864928, 0.11195961, 0.06613588, ..., 0.14300013, 0.15073101,
              0.06193411],
             [0.10042011, 0.08104166, 0.10368333, ..., 0.10997637, 0.10437962,
              0.09832105]], dtype=float32)

    • p_outlier
      (chain, draw)
      float32
      0.00135 0.005336 ... 0.008597

      array([[0.0013504 , 0.00533634, 0.00458578, ..., 0.01001942, 0.01956319,
              0.00332052],
             [0.00673035, 0.00346698, 0.0291583 , ..., 0.00110085, 0.00737388,
              0.00859677]], dtype=float32)

    • a
      (chain, draw)
      float32
      1.573 1.524 1.579 ... 1.546 1.563

      array([[1.5730585, 1.5237738, 1.5790713, ..., 1.4876293, 1.4824865,
              1.6106321],
             [1.5608598, 1.5764627, 1.5554554, ..., 1.5537058, 1.5459083,
              1.5630828]], dtype=float32)

  • Indexes: (2)
    • chain
      PandasIndex

      PandasIndex(Index([0, 1], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
             ...
             990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
            dtype='int64', name='draw', length=1000))

  • Attributes: (8)

    created_at :

    2024-09-18T23:33:36.552645+00:00

    arviz_version :

    0.19.0

    inference_library :

    pymc

    inference_library_version :

    5.16.2

    sampling_time :

    30.324628829956055

    tuning_steps :

    1000

    modeling_interface :

    bambi

    modeling_interface_version :

    0.14.0

  
  
• log_likelihood

  <xarray.Dataset> Size: 16MB
  Dimensions:      (chain: 2, draw: 1000, __obs__: 1000)
  Coordinates:
    * chain        (chain) int64 16B 0 1
    * draw         (draw) int64 8kB 0 1 2 3 4 5 6 ... 993 994 995 996 997 998 999
    * __obs__      (__obs__) int64 8kB 0 1 2 3 4 5 6 ... 994 995 996 997 998 999
  Data variables:
      rt,response  (chain, draw, __obs__) float64 16MB -3.266 -2.511 ... -1.498
  Attributes:
      modeling_interface:          bambi
      modeling_interface_version:  0.14.0

  xarray.Dataset

  • Dimensions:
    • chain: 2
    • draw: 1000
    • __obs__: 1000
  • Coordinates: (3)
    • chain
      (chain)
      int64
      0 1

      array([0, 1])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

    • __obs__
      (__obs__)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

  • Data variables: (1)
    • rt,response
      (chain, draw, __obs__)
      float64
      -3.266 -2.511 ... -2.301 -1.498

      array([[[-3.26646071, -2.51058283, -3.01556917, ..., -1.20096499,
               -2.30571449, -1.52950116],
              [-3.37844241, -2.57102697, -3.00242304, ..., -1.08399074,
               -2.34722976, -1.49929992],
              [-3.48744945, -2.49185328, -2.87280863, ..., -1.20119257,
               -2.2744762 , -1.46338872],
              ...,
              [-3.37864152, -2.58215128, -2.77360277, ..., -1.19353063,
               -2.34643053, -1.46336375],
              [-3.33108289, -2.62137536, -3.04579322, ..., -1.06662279,
               -2.390417  , -1.51019722],
              [-3.39928476, -2.46272643, -3.01741035, ..., -1.24008603,
               -2.25895354, -1.49336871]],
      
             [[-3.47210119, -2.49624285, -2.90540586, ..., -1.24057575,
               -2.27416949, -1.44574064],
              [-3.43716243, -2.50843074, -3.059324  , ..., -1.14381449,
               -2.29433404, -1.48588919],
              [-3.31312547, -2.50342222, -2.80930747, ..., -1.32665319,
               -2.29230635, -1.50103593],
              ...,
              [-3.2788995 , -2.51551823, -2.98522557, ..., -1.22132919,
               -2.30430656, -1.50731197],
              [-3.4666517 , -2.52031837, -2.89630192, ..., -1.19801935,
               -2.29509873, -1.45261314],
              [-3.35673096, -2.51353533, -3.01904874, ..., -1.19021741,
               -2.30098352, -1.49794317]]])

  • Indexes: (3)
    • chain
      PandasIndex

      PandasIndex(Index([0, 1], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
             ...
             990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
            dtype='int64', name='draw', length=1000))

    • __obs__
      PandasIndex

      PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
             ...
             990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
            dtype='int64', name='__obs__', length=1000))

  • Attributes: (2)

    modeling_interface :

    bambi

    modeling_interface_version :

    0.14.0

  
  
• sample_stats

  <xarray.Dataset> Size: 252kB
  Dimensions:                (chain: 2, draw: 1000)
  Coordinates:
    * chain                  (chain) int64 16B 0 1
    * draw                   (draw) int64 8kB 0 1 2 3 4 5 ... 995 996 997 998 999
  Data variables: (12/17)
      acceptance_rate        (chain, draw) float64 16kB 0.9887 0.62 ... 0.9326
      diverging              (chain, draw) bool 2kB False False ... False False
      energy                 (chain, draw) float64 16kB 2.021e+03 ... 2.014e+03
      energy_error           (chain, draw) float64 16kB -0.1611 ... 0.05626
      index_in_trajectory    (chain, draw) int64 16kB 6 -3 -6 4 6 ... -2 -5 4 5 5
      largest_eigval         (chain, draw) float64 16kB nan nan nan ... nan nan
      ...                     ...
      process_time_diff      (chain, draw) float64 16kB 0.01018 ... 0.01997
      reached_max_treedepth  (chain, draw) bool 2kB False False ... False False
      smallest_eigval        (chain, draw) float64 16kB nan nan nan ... nan nan
      step_size              (chain, draw) float64 16kB 0.5852 0.5852 ... 0.461
      step_size_bar          (chain, draw) float64 16kB 0.5291 0.5291 ... 0.5085
      tree_depth             (chain, draw) int64 16kB 3 3 4 4 3 3 ... 3 3 3 3 4 4
  Attributes:
      created_at:                  2024-09-18T23:33:36.567622+00:00
      arviz_version:               0.19.0
      inference_library:           pymc
      inference_library_version:   5.16.2
      sampling_time:               30.324628829956055
      tuning_steps:                1000
      modeling_interface:          bambi
      modeling_interface_version:  0.14.0

  xarray.Dataset

  • Dimensions:
    • chain: 2
    • draw: 1000
  • Coordinates: (2)
    • chain
      (chain)
      int64
      0 1

      array([0, 1])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

  • Data variables: (17)
    • acceptance_rate
      (chain, draw)
      float64
      0.9887 0.62 ... 0.7555 0.9326

      array([[0.98869794, 0.62000823, 0.92609966, ..., 0.95851026, 0.88880842,
              0.11423812],
             [0.96251723, 0.82172299, 0.62412233, ..., 0.78891661, 0.75548026,
              0.93263182]])

    • diverging
      (chain, draw)
      bool
      False False False ... False False

      array([[False, False, False, ..., False, False, False],
             [False, False, False, ..., False, False, False]])

    • energy
      (chain, draw)
      float64
      2.021e+03 2.022e+03 ... 2.014e+03

      array([[2021.4846039 , 2022.11065874, 2017.68579753, ..., 2018.04684843,
              2020.45976946, 2023.51311413],
             [2016.12159994, 2016.59242019, 2020.61557583, ..., 2016.08280258,
              2016.93678484, 2014.06006628]])

    • energy_error
      (chain, draw)
      float64
      -0.1611 -0.1302 ... -0.1495 0.05626

      array([[-0.16113148, -0.13018934, -0.03592278, ...,  0.00613487,
              -0.45987121,  1.99731273],
             [-0.13396856,  0.30559856,  1.32401175, ...,  0.34430895,
              -0.14951642,  0.0562646 ]])

    • index_in_trajectory
      (chain, draw)
      int64
      6 -3 -6 4 6 -5 5 ... 0 -2 -5 4 5 5

      array([[ 6, -3, -6, ..., -3,  3, -5],
             [ 3,  4,  2, ...,  4,  5,  5]])

    • largest_eigval
      (chain, draw)
      float64
      nan nan nan nan ... nan nan nan nan

      array([[nan, nan, nan, ..., nan, nan, nan],
             [nan, nan, nan, ..., nan, nan, nan]])

    • lp
      (chain, draw)
      float64
      -2.017e+03 ... -2.013e+03

      array([[-2017.44194748, -2013.59115992, -2014.4453071 , ...,
              -2016.2303492 , -2016.65226505, -2016.84925906],
             [-2012.99596456, -2014.1546112 , -2018.77659385, ...,
              -2015.7061619 , -2012.64406166, -2013.2134788 ]])

    • max_energy_error
      (chain, draw)
      float64
      -0.4445 1.175 ... 0.5667 0.1641

      array([[-0.44450216,  1.17530501,  0.33833815, ...,  0.16117841,
               0.58204568,  3.91370518],
             [ 0.26394248,  0.45347987,  1.56880813, ...,  0.34430895,
               0.56674545,  0.16409927]])

    • n_steps
      (chain, draw)
      float64
      7.0 7.0 11.0 11.0 ... 7.0 15.0 15.0

      array([[ 7.,  7., 11., ...,  7.,  7.,  7.],
             [ 7.,  7.,  7., ...,  7., 15., 15.]])

    • perf_counter_diff
      (chain, draw)
      float64
      0.01018 0.01004 ... 0.01994 0.01997

      array([[0.01018211, 0.01004128, 0.01499515, ..., 0.00994783, 0.00998986,
              0.00998649],
             [0.0098437 , 0.0119226 , 0.00984869, ..., 0.01024293, 0.01993772,
              0.0199689 ]])

    • perf_counter_start
      (chain, draw)
      float64
      298.5 298.5 298.5 ... 308.0 308.0

      array([[298.4693637 , 298.47972678, 298.48992971, ..., 309.9416699 ,
              309.95178245, 309.96193316],
             [297.63047153, 297.64047942, 297.6525782 , ..., 307.99174811,
              308.00216422, 308.02226123]])

    • process_time_diff
      (chain, draw)
      float64
      0.01018 0.01004 ... 0.01994 0.01997

      array([[0.01018197, 0.01004121, 0.01499523, ..., 0.00994818, 0.00998178,
              0.00998675],
             [0.00984393, 0.01192354, 0.00984902, ..., 0.01024296, 0.01993768,
              0.01996913]])

    • reached_max_treedepth
      (chain, draw)
      bool
      False False False ... False False

      array([[False, False, False, ..., False, False, False],
             [False, False, False, ..., False, False, False]])

    • smallest_eigval
      (chain, draw)
      float64
      nan nan nan nan ... nan nan nan nan

      array([[nan, nan, nan, ..., nan, nan, nan],
             [nan, nan, nan, ..., nan, nan, nan]])

    • step_size
      (chain, draw)
      float64
      0.5852 0.5852 ... 0.461 0.461

      array([[0.58515462, 0.58515462, 0.58515462, ..., 0.58515462, 0.58515462,
              0.58515462],
             [0.46100977, 0.46100977, 0.46100977, ..., 0.46100977, 0.46100977,
              0.46100977]])

    • step_size_bar
      (chain, draw)
      float64
      0.5291 0.5291 ... 0.5085 0.5085

      array([[0.52914611, 0.52914611, 0.52914611, ..., 0.52914611, 0.52914611,
              0.52914611],
             [0.50846335, 0.50846335, 0.50846335, ..., 0.50846335, 0.50846335,
              0.50846335]])

    • tree_depth
      (chain, draw)
      int64
      3 3 4 4 3 3 4 3 ... 2 3 3 3 3 3 4 4

      array([[3, 3, 4, ..., 3, 3, 3],
             [3, 3, 3, ..., 3, 4, 4]])

  • Indexes: (2)
    • chain
      PandasIndex

      PandasIndex(Index([0, 1], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
             ...
             990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
            dtype='int64', name='draw', length=1000))

  • Attributes: (8)

    created_at :

    2024-09-18T23:33:36.567622+00:00

    arviz_version :

    0.19.0

    inference_library :

    pymc

    inference_library_version :

    5.16.2

    sampling_time :

    30.324628829956055

    tuning_steps :

    1000

    modeling_interface :

    bambi

    modeling_interface_version :

    0.14.0

  
  
• observed_data

  <xarray.Dataset> Size: 16kB
  Dimensions:                  (__obs__: 1000, rt,response_extra_dim_0: 2)
  Coordinates:
    * __obs__                  (__obs__) int64 8kB 0 1 2 3 4 ... 996 997 998 999
    * rt,response_extra_dim_0  (rt,response_extra_dim_0) int64 16B 0 1
  Data variables:
      rt,response              (__obs__, rt,response_extra_dim_0) float32 8kB 2...
  Attributes:
      created_at:                  2024-09-18T23:33:36.572500+00:00
      arviz_version:               0.19.0
      inference_library:           pymc
      inference_library_version:   5.16.2
      modeling_interface:          bambi
      modeling_interface_version:  0.14.0

  xarray.Dataset

  • Dimensions:
    • __obs__: 1000
    • rt,response_extra_dim_0: 2
  • Coordinates: (2)
    • __obs__
      (__obs__)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

    • rt,response_extra_dim_0
      (rt,response_extra_dim_0)
      int64
      0 1

      array([0, 1])

  • Data variables: (1)
    • rt,response
      (__obs__, rt,response_extra_dim_0)
      float32
      2.447 -1.0 3.504 ... 1.0 1.869 1.0

      array([[ 2.4472353 , -1.        ],
             [ 3.5043278 ,  1.        ],
             [ 0.50780696, -1.        ],
             ...,
             [ 0.56721747,  1.        ],
             [ 3.1607578 ,  1.        ],
             [ 1.869449  ,  1.        ]], dtype=float32)

  • Indexes: (2)
    • __obs__
      PandasIndex

      PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
             ...
             990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
            dtype='int64', name='__obs__', length=1000))

    • rt,response_extra_dim_0
      PandasIndex

      PandasIndex(Index([0, 1], dtype='int64', name='rt,response_extra_dim_0'))

  • Attributes: (6)

    created_at :

    2024-09-18T23:33:36.572500+00:00

    arviz_version :

    0.19.0

    inference_library :

    pymc

    inference_library_version :

    5.16.2

    modeling_interface :

    bambi

    modeling_interface_version :

    0.14.0

  
  

In [7]:

ddm_model_lapse.summary(lapse_trace)

ddm_model_lapse.summary(lapse_trace)

Out[7]:

	mean	sd	hdi_3%	hdi_97%	mcse_mean	mcse_sd	ess_bulk	ess_tail	r_hat
v	0.513	0.030	0.455	0.570	0.001	0.001	1129.0	1206.0	1.0
a	1.552	0.028	1.496	1.602	0.001	0.001	1418.0	1211.0	1.0
p_outlier	0.008	0.007	0.000	0.021	0.000	0.000	1182.0	907.0	1.0
t	0.102	0.021	0.064	0.143	0.001	0.000	1195.0	1247.0	1.0
z	0.502	0.013	0.479	0.526	0.000	0.000	987.0	1264.0	1.0

In [8]:

ddm_model_lapse.plot_trace(lapse_trace)
plt.tight_layout()

ddm_model_lapse.plot_trace(lapse_trace) plt.tight_layout()

Disable lapse probabilities and distributions¶

When p_outliers is set to None or 0, lapse probability and distributions will be ignored. They will not be included in the model output.

Note

If `p_outlier` is set to `None` or 0 but `lapse` is not set to None, you will receive a warning. Please remember to also set `lapse` to `None`.

In [9]:

ddm_model_no_lapse = hssm.HSSM(data=ddm_data, p_outlier=None, lapse=None)
ddm_model_no_lapse

ddm_model_no_lapse = hssm.HSSM(data=ddm_data, p_outlier=None, lapse=None) ddm_model_no_lapse

Model initialized successfully.

Out[9]:

Hierarchical Sequential Sampling Model
Model: ddm

Response variable: rt,response
Likelihood: analytical
Observations: 1000

Parameters:

v:
    Prior: Normal(mu: 0.0, sigma: 2.0)
    Explicit bounds: (-inf, inf)

a:
    Prior: HalfNormal(sigma: 2.0)
    Explicit bounds: (0.0, inf)

z:
    Prior: Uniform(lower: 0.0, upper: 1.0)
    Explicit bounds: (0.0, 1.0)

t:
    Prior: HalfNormal(sigma: 2.0)
    Explicit bounds: (0.0, inf)

In [10]:

ddm_model_no_lapse_warn = hssm.HSSM(data=ddm_data, p_outlier=0)
ddm_model_no_lapse_warn

ddm_model_no_lapse_warn = hssm.HSSM(data=ddm_data, p_outlier=0) ddm_model_no_lapse_warn

You have specified the `lapse` argument to include a lapse distribution, but `p_outlier` is set to either 0 or None. Your lapse distribution will be ignored.

Model initialized successfully.

Out[10]:

Hierarchical Sequential Sampling Model
Model: ddm

Response variable: rt,response
Likelihood: analytical
Observations: 1000

Parameters:

v:
    Prior: Normal(mu: 0.0, sigma: 2.0)
    Explicit bounds: (-inf, inf)

a:
    Prior: HalfNormal(sigma: 2.0)
    Explicit bounds: (0.0, inf)

z:
    Prior: Uniform(lower: 0.0, upper: 1.0)
    Explicit bounds: (0.0, 1.0)

t:
    Prior: HalfNormal(sigma: 2.0)
    Explicit bounds: (0.0, inf)