Basic PVT Setup and Example

Import AgentClient from module ia.gaius.agent_client

[1]:

from ia.gaius.agent_client import AgentClient

[2]:

agent_info = {'name': '',
              'domain': 'gaius-api',
              'secure': False,
              'api_key': 'ABCD-1234'}

[3]:

agent = AgentClient(agent_info)

[4]:

agent.connect()

[4]:

{'connection': 'okay', 'agent': 'simple'}

[5]:

agent.show_status()

[5]:

{'AUTOLEARN': False,
 'PREDICT': True,
 'SLEEPING': False,
 'emotives': {},
 'last_learned_model_name': 'cda375bca8ec4af85dc666ccff417e4841a1b475',
 'models_kb': '{KB| objects: 150}',
 'name': 'P1',
 'size_WM': 0,
 'target': '',
 'time': 300,
 'vectors_kb': '{KB| objects: 149}'}

[6]:

agent.get_all_genes(nodes=['P1'])

[6]:

{'genes': {'always_update_frequencies': False,
  'auto_learn_algorithm': 'basic',
  'auto_learn_metric': 'WM_size',
  'classifier': 'CVC',
  'max_predictions': 100,
  'near_vector_count': 3,
  'persistence': 5,
  'predict_on_nth_event': 1,
  'quiescence': 3,
  'recall_threshold': 0.2,
  'smoothness': 3,
  'sort_metric': 'potential'}}

[ ]:

Iris Dataset Training

Train and test on iris-flowers dataset using the PerformanceValidationTest class from ia.gaius.pvt. This dataset is stored locally in GDF format.

[7]:

from ia.gaius.pvt import PerformanceValidationTest

Configure the test to:

Conduct a classification test
Train using an 80/20 training/testing data split
Use 100% of the dataset (150 records)
learn after guessing during training phase (test_prediction_strategy = continuous)
Clear All Memory on agent before conducting PVT
Shuffle the data so it is in random order

[8]:

pvt = PerformanceValidationTest(agent=agent, ingress_nodes=['P1'], query_nodes=['P1'], num_of_tests=1, pct_of_ds=100, pct_res_4_train=80,
                                     test_type='classification', dataset_location='filepath', results_filepath=None,
                                     ds_filepath='../../../../datasets/iris-flowers/', test_prediction_strategy='continuous',
                                     clear_all_memory_before_training=True, turn_prediction_off_during_training=False, shuffle=True)

num_of_tests      = 1

ds_filepath       = ../../../../datasets/iris-flowers/

pct_of_ds         = 100

pct_res_4_train   = 80

summarize_for_single_node status   = False

[9]:

pvt.conduct_pvt()

Conducting Classification PVT...

Conducting Test # 0

---------------------

Length of Training Set = 120

Length of Testing Set  = 30

Clearing memory of selected ingress nodes...
Initialize labels set...
{'P1': set()}
Created labels set...
Preparing to train agent...
train - 0
train - 100
Finished training agent!
test - 0
Getting Classification Metrics...
Saving results to pvt_results...
Plotting Results...
-----------------Test#0-P1-Plots-----------------

-----------------Test#0-hive-Plots-----------------