# user configuration
delim = ','
disc_type = "static"
task_name = "discovery"
max_len = 5
beamsize = 100
ncutpoints = 5
target_type = "numeric"
results = ""
datasetnames = ["baseball","autoMPG8","dee","ele-1","forestFires","concrete",\
"treasury","wizmir","abalone","puma32h","ailerons","elevators",\
"bikesharing","california","house"]
for datasetname in datasetnames:
print("Dataset: " + datasetname)
filename = "./data/numeric target/" + datasetname + ".csv"
df = pd.read_csv(filename, delimiter=delim)
model = SSDC(target_type,
max_depth=max_len,
beam_width=beamsize,
iterative_beam_width=1,
n_cutpoints=ncutpoints,
task=task_name,
discretization=disc_type,
gain="normalized")
model.fit(df)
# fit to new data
model.measures["nsamples_train"] = len(df)
results = attach_results(model, results, datasetname)
print2folder(model, results, "SSDpp")
# -*- coding: utf-8 -*-
import pandas as pd
import sys
sys.path
from src.util.results2folder import attach_results, print2folder
from _classes import SSDC
# load data
datasetname = "baseball"
delim = ','
filename = "./data/numeric target/" + datasetname + ".csv"
df = pd.read_csv(filename, delimiter=delim)
# user configuration
task_name = "discovery"
target_type = "numeric"
# load class and fit to data
model = SSDC(task=task_name)
model.fit(df)
#print("model measures : " +str(model.measures) + "\n")
print(model)
disc_type = "static"
task_name = "discovery"
max_len = 5
beamsize = 100
ncutpoints = 5
target_type = "numeric"
# load data
filename = "./data/application/" + datasetname + ".csv"
df = pd.read_csv(filename, delimiter=delim)
# load model
model = SSDC(target_type,
max_depth=max_len,
beam_width=beamsize,
iterative_beam_width=1,
n_cutpoints=ncutpoints,
task=task_name,
discretization=disc_type,
gain="normalized",
max_rules=4)
model.fit(df)
print(model)
#ruleset = model.rule_sets
#overlap = []
#subgroup_sets_support= [set(rset) for rset in ruleset]
#rules_usg = []
#subgroup_sets_usage= []
#statistic_rules = []
#for r in range(len(subgroup_sets_support)):
# previous_sets = [subgroup_sets_support[ii] for ii in range(r)]
beamsize = 100
ncutpoints = 5
iterative = 1
target_type = "numeric"
maxdepth_list = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
datasetnames = ["baseball","autoMPG8","dee","ele-1","forestFires","concrete",\
"treasury","wizmir","abalone","puma32h","ailerons","elevators",\
"bikesharing","california","house"]
results = ""
for datasetname in datasetnames:
print("Dataset name: " + datasetname)
for maxdepth in maxdepth_list:
print(" maximum depth of search: " + str(maxdepth))
filename = "./data/numeric target/" + datasetname + ".csv"
df = pd.read_csv(filename, delimiter=delim)
model = SSDC(target_type,
max_depth=maxdepth,
beam_width=beamsize,
iterative_beam_width=iterative,
n_cutpoints=ncutpoints,
task=task_name,
discretization=disc_type,
gain="normalized")
model.fit(df)
model.measures["maxdepth"] = maxdepth
results = attach_results(model, results, datasetname)
results = results.rstrip(", \n")
print2folder(model, results, "maxdepth_results")