Configuration Parameters Reference¶
Here are the configuration parameters supported by pebl. These can set in a configuration file or via the config.set() function.
data¶
-
data.discretize
¶ Number of bins used to discretize data. Specify 0 to indicate that data should not be discretized. default=0
-
data.filename
¶ File to read data from. default=None
-
data.text
¶ The text of a dataset included in config file. default=
learner¶
-
learner.numtasks
¶ Number of learner tasks to run. default=1
-
learner.type
¶ - Type of learner to use.
- The following learners are included with pebl:
- greedy.GreedyLearner
- simanneal.SimulatedAnnealingLearner
- exhaustive.ListLearner
default=greedy.GreedyLearner
greedy¶
-
greedy.max_iterations
¶ Maximum number of iterations to run. default=1000
-
greedy.max_time
¶ Maximum learner runtime in seconds. default=0
-
greedy.max_unimproved_iterations
¶ Maximum number of iterations without score improvement before a restart. default=500
-
greedy.seed
¶ Starting network for a greedy search. default=
simanneal¶
-
simanneal.delta_temp
¶ Change in temp between steps. default=0.5
-
simanneal.max_iters_at_temp
¶ Max iterations at any temperature. default=100
-
simanneal.seed
¶ Starting network for a greedy search. default=
-
simanneal.start_temp
¶ Starting temperature for a run. default=100.0
localscore_cache¶
-
localscore_cache.maxsize
¶ Max number of localscores to cache. Default=-1 means unlimited size. default=-1
result¶
-
result.filename
¶ The name of the result output file default=result.pebl
-
result.format
¶ The format for the pebl result file (pickle or html). default=pickle
-
result.size
¶ Number of top-scoring networks to save. Specify 0 to indicate that all scored networks should be saved. default=1000
-
gibbs.burnin
¶ Burn-in period for the gibbs sampler (specified as a multiple of the number of missing values) default=10
-
gibbs.max_iterations
¶ Stopping criteria for the gibbs sampler.
The number of Gibb’s sampler iterations to run. Should be a valid python expression using the variable n (number of missing values). Examples:
- n**2 (for n-squared iterations)
- 100 (for 100 iterations)
default=n**2
multiprocess¶
-
multiprocess.poolsize
¶ Number of processes to run concurrently (0 means no limit) default=0
xgrid¶
-
xgrid.controller
¶ Hostname or IP of the Xgrid controller. default=
-
xgrid.grid
¶ Id of the grid to use at the Xgrid controller. default=0
-
xgrid.password
¶ Password for the Xgrid controller. default=
-
xgrid.peblpath
¶ Full path to the pebl script on Xgrid agents default=pebl
-
xgrid.pollinterval
¶ Time (in secs) to wait between polling the Xgrid controller. default=60.0
ipython1¶
-
ipython1.controller
¶ IPython1 TaskController (default is 127.0.0.1:10113) default=127.0.0.1:10113
ec2¶
-
ec2.config
¶ EC2 config file. This is kept seperate from pebl config because it contains authentication keys, etc. default=
-
ec2.max_count
¶ Maximum number of EC2 instances to create (default=0 means the same number as ec2.min_count). default=0
-
ec2.min_count
¶ Minimum number of EC2 instances to create (default=1). default=1