mmerce

from bigml.api import BigML, FINISHED
api = BigML()
DATE = '2013-10-05T00:25:36.973000'
PAGE_LENGTH = 20 # number of resources to be retrieved in one list call
# q_w: query_string with the query parameters. In this case, creation data and
# maximum number of resources to be returned.
QUERY_STRING = "created__gte=%s" % DATE

# api_function: function that should be used to retrieve the resources. In this
# case we use the list_sources call to retrieve sources. We could use any

mmerce

bigmler analyze --features \
                --dataset dataset/535f5ecc37203f272e0001b4

mmerce

bigmler analyze --features \
                --dataset dataset/535f5ecc37203f272e0001b4 \
                --penalty 0.002 --staleness 3 --k-folds 2

mmerce

bigmler analyze --features \
                --dataset dataset/535f5ecc37203f272e0001b4 \
                --k-folds 2 --max-parallel-models 2 \
                --max-parallel-evaluations 2

mmerce

bigmler analyze --features \
                --dataset dataset/535f5ecc37203f272e0001b4 \
                --maximize recall

mmerce

bigmler analyze --nodes \
                --dataset dataset/535f5ecc37203f272e0001b4

mmerce

bigmler analyze --nodes \
                --dataset dataset/535f5ecc37203f272e0001b4 \
                --min-nodes 5 --max-nodes 25 --nodes-step 5 \
                --k-folds 2 --maximize phi

mmerce

bigmler analyze --dataset dataset/536653050af5e86d9c01549e \
                --cross-validation --k-folds 3

mmerce

Node.js example: local predictions

Using the node.js bindings and its local model object to predict from a csv file.

Usage:

cat input.csv | node act_model.js > predictions.csv

mmerce

Python example: local predictions

Using the python bindings and its local model object to predict from a csv file.

Usage:

cat input.csv | python act_model.js > predictions.csv