GraphMl belief_propagation

---

ml.belief_propagation(self, prior_property, posterior_property, edge_weight_property=None, convergence_threshold=None, max_iterations=None)

Classification on sparse data using Belief Propagation.

Parameters:

prior_property : unicode

Name of the vertex property which contains the prior belief for the vertex.

posterior_property : unicode

Name of the vertex property which will contain the posterior belief for each vertex.

edge_weight_property : unicode (default=None)

Name of the edge property that contains the edge weight for each edge.

convergence_threshold : float64 (default=None)

Belief propagation will terminate when the average change in posterior beliefs between supersteps is less than or equal to this threshold.

max_iterations : int32 (default=None)

The maximum number of supersteps that the algorithm will execute. The valid range is all positive int.

Returns:

: dict

Progress report for belief propagation in the format of a multiple-line string.

Belief propagation by the sum-product algorithm. This algorithm analyzes a graphical model with prior beliefs using sum product message passing. The priors are read from a property in the graph, the posteriors are written to another property in the graph. This is the GraphX-based implementation of belief propagation.

See Loopy Belief Propagation for a more in-depth discussion of BP and LBP.

Examples

>>> graph.ml.belief_propagation("value", "lbp_output", string_output = True, state_space_size = 5, max_iterations = 6)

{u'log': u'Vertex Count: 80000\nEdge Count: 318398\nAtkPregel engine has completed iteration 1  The average delta is 0.6853413553663811\nAtkPregel engine has completed iteration 2  The average delta is 0.38626944467366386\nAtkPregel engine has completed iteration 3  The average delta is 0.2365329376479823\nAtkPregel engine has completed iteration 4  The average delta is 0.14170840479478952\nAtkPregel engine has completed iteration 5  The average delta is 0.08676093923623975\n', u'time': 70.248999999999995}

>>> graph.query.gremlin("g.V [0..4]")

{u'results': [{u'vertex_type': u'VA', u'target': 12779523, u'lbp_output': u'0.9485759073302487, 0.001314151524421738, 0.040916996746627056, 0.001397331576080859, 0.0077956128226217315', u'_type': u'vertex', u'value': u'0.125 0.125 0.5 0.125 0.125', u'titanPhysicalId': 4, u'_id': 4}, {u'vertex_type': u'VA', u'titanPhysicalId': 8, u'lbp_output': u'0.7476996339617544, 0.0021769696832380173, 0.24559940461433935, 0.0023272253558738786, 0.002196766384794168', u'_type': u'vertex', u'value': u'0.125 0.125 0.5 0.125 0.125', u'source': 7798852, u'_id': 8}, {u'vertex_type': u'TR', u'target': 13041863, u'lbp_output': u'0.7288360734608738, 0.07162637515155296, 0.15391773902131053, 0.022620779563724287, 0.02299903280253846', u'_type': u'vertex', u'value': u'0.5 0.125 0.125 0.125 0.125', u'titanPhysicalId': 12, u'_id': 12}, {u'vertex_type': u'TR', u'titanPhysicalId': 16, u'lbp_output': u'0.9996400056392905, 9.382190989071985E-5, 8.879762476576982E-5, 8.867586165695348E-5, 8.869896439624652E-5', u'_type': u'vertex', u'value': u'0.5 0.125 0.125 0.125 0.125', u'source': 11731127, u'_id': 16}, {u'vertex_type': u'TE', u'titanPhysicalId': 20, u'lbp_output': u'0.004051247779081896, 0.2257641948616088, 0.01794622866204068, 0.7481547408142287, 0.004083587883039745', u'_type': u'vertex', u'value': u'0.125 0.125 0.5 0.125 0.125', u'source': 3408035, u'_id': 20}], u'run_time_seconds': 1.042}