from suanpan.docker.arguments import Int, String, Table, Bool, Float, ListOfString
from arguments import SklearnModel
from catboost import CatBoostRegressor
@dc.input(
Table(
key="inputData", table="inputTable", partition="inputPartition", required=True
)
)
@dc.column(ListOfString(key="featureColumns", default=[]))
@dc.column(String(key="labelColumn", default="MEDV"))
@dc.param(
Int(
key="iterations",
default=1000,
help="The maximum number of trees that can be built when solving machine learning problems.",
)
)
@dc.param(Float(key="learningRate", default=0.03, help="The learning rate."))
@dc.param(Int(key="depth", default=6, help="Depth of the tree."))
@dc.param(
Float(
key="l2LeafReg",
default=3.0,
help="Coefficient at the L2 regularization term of the cost function.",
)
)
@dc.param(
Float(
key="rsm",
String(
key="missing",
default="none",
help="Available options are ‘none’, ‘drop’, and ‘raise’.",
)
)
@dc.param(
String(
key="method",
default="lbfgs",
help="‘newton’, ‘bfgs’, ‘lbfgs’, ‘powell’, ‘cg’, ‘ncg’, ‘basinhopping’,"
" ‘minimize’",
)
)
@dc.param(
Int(key="maxiter", default=35, help="The maximum number of iterations to perform.")
)
@dc.param(Int(key="disp", default=1, help="Set to True to print convergence messages."))
@dc.output(SklearnModel(key="outputModel"))
def SPLogit(context):
# 从 Context 中获取相关数据
args = context.args
# 查看上一节点发送的 args.inputData 数据
df = args.inputData
featureColumns = args.featureColumns
labelColumn = args.labelColumn
features = df[featureColumns].values
label = df[labelColumn].values
String(
key="missing",
default="none",
help="Available options are ‘none’, ‘drop’, and ‘raise’.",
))
@dc.param(
String(
key="trend",
default="c",
help=
"Whether to include a constant or not. ‘c’ includes constant, ‘nc’ no constant.",
))
@dc.param(String(key="method", default="cmle", help="‘cmle’, ‘mle’"))
@dc.param(
Int(key="maxiter",
default=35,
help="The maximum number of function evaluations."))
@dc.param(
Int(key="disp",
default=1,
help="If True, convergence information is output."))
@dc.param(
Int(
key="maxlag",
default=None,
help="If ic is None, then maxlag is the lag length used in fit.",
))
@dc.output(SklearnModel(key="outputModel"))
def SPAR(context):
# 从 Context 中获取相关数据
args = context.args
@dc.param(
String(
key="trend",
default="c",
help="Whether to include a constant or not. ‘c’ includes constant, ‘nc’ no constant.",
)
)
@dc.param(
String(
key="method",
default="css-mle",
help="This is the loglikelihood to maximize.‘css-mle’,’mle’,’css’",
)
)
@dc.param(
Int(key="maxiter", default=500, help="The maximum number of function evaluations.")
)
@dc.param(
Int(key="disp", default=5, help="If True, convergence information is printed.")
)
@dc.output(SklearnModel(key="outputModel"))
def SPARMA(context):
# 从 Context 中获取相关数据
args = context.args
# 查看上一节点发送的 args.inputData 数据
inputdata = args.inputData
inputdata = (
pd.DataFrame(inputdata[args.labelColumn].values, index=inputdata.index)
if args.timestampIndex
else pd.DataFrame(
inputdata[args.labelColumn].values,
from suanpan.docker.arguments import Int, String, Bool, Float, ListOfString, Table
import lightgbm as lgb
from arguments import SklearnModel
@dc.input(
Table(key="inputData",
table="inputTable",
partition="inputPartition",
required=True))
@dc.column(ListOfString(key="featureColumns", default=["f1", "f2", "f3",
"f4"]))
@dc.column(String(key="labelColumn", default="label"))
@dc.param(
Int(key="maxDepth",
default=-1,
help="Maximum tree depth for base learners"))
@dc.param(
String(
key="boostingType",
default="gbdt",
help="Specify which booster to use: 'goss', 'rf' or 'dart'",
))
@dc.param(
Int(key="numLeaves",
default=31,
help="Maximum tree leaves for base learners."))
@dc.param(
Float(key="learningRate", default=0.1, help="Boosting learning rate."))
@dc.param(
Int(key="nEstimators", default=100,
@dc.param(
ListOfFloat(
key="ar",
default=[0.75, -0.25],
help=
"coefficient for autoregressive lag polynomial, including zero lag",
))
@dc.param(
ListOfFloat(
key="ma",
default=[0.65, 0.35],
help=
"coefficient for moving-average lag polynomial, including zero lag",
))
@dc.param(
Int(key="nsample", default=250, help="length of simulated time series"))
@dc.param(Float(key="sigma", default=1.0, help="standard deviation of noise"))
@dc.param(Int(key="randomSeed", default=12345, help="random seed"))
@dc.param(Bool(key="dateCol", default=True, help="date in dataset"))
@dc.param(
String(
key="startDate",
default="19800131",
help="The first abbreviated date, for instance, '1965q1' or '1965m1'",
))
@dc.param(String(key="freq", default="M", help="DateOffset"))
@dc.output(Csv(key="outputData"))
def SPARMASample(context):
# 从 Context 中获取相关数据
args = context.args
# 查看上一节点发送的 args.inputData 数据
# coding=utf-8
from __future__ import absolute_import, print_function
from suanpan.docker import DockerComponent as dc
from suanpan.docker.arguments import Csv, ListOfString, String, Int
import statsmodels.api as sm
from arguments import SklearnModel
@dc.input(Csv(key="inputData"))
@dc.column(ListOfString(key="featureColumns", default=["a", "b", "c", "d"]))
@dc.column(String(key="labelColumn", default="e"))
@dc.param(Int(key="rho", default=1, help="Order of the autoregressive covariance"))
@dc.param(
String(
key="missing",
default="none",
help="Available options are ‘none’, ‘drop’, and ‘raise’.",
)
)
@dc.param(String(key="method", default="pinv", help="Can be “pinv”, “qr”. "))
@dc.output(SklearnModel(key="outputModel"))
def SPGLSAR(context):
# 从 Context 中获取相关数据
args = context.args
# 查看上一节点发送的 args.inputData 数据
df = args.inputData
featureColumns = args.featureColumns
labelColumn = args.labelColumn
@dc.column(String(key="labelColumn", default="e"))
@dc.param(
String(
key="family",
default="Gaussian",
help=
"The default is Gaussian. Binomial, Gamma, Gaussian, InverseGaussian"
"NegativeBinomial, Poisson, Tweedie",
))
@dc.param(
String(
key="missing",
default="none",
help="Available options are ‘none’, ‘drop’, and ‘raise’.",
))
@dc.param(Int(key="maxiter", default=100, help="Default is 100."))
@dc.output(SklearnModel(key="outputModel"))
def SPGLM(context):
# 从 Context 中获取相关数据
args = context.args
# 查看上一节点发送的 args.inputData 数据
df = args.inputData
featureColumns = args.featureColumns
labelColumn = args.labelColumn
features = df[featureColumns].values
label = df[labelColumn].values
family = args.family
result = getattr(sm.families, family)()
String(
key="M",
default="HuberT",
help=
"The default is LeastSquares. HuberT, RamsayE, AndrewWave, TrimmedMean"
"Hampel, TukeyBiweight",
))
@dc.param(
String(
key="missing",
default="none",
help="Available options are ‘none’, ‘drop’, and ‘raise’.",
))
@dc.param(
Int(key="maxiter",
default=50,
help="The maximum number of iterations to try."))
@dc.output(SklearnModel(key="outputModel"))
def SPRLM(context):
# 从 Context 中获取相关数据
args = context.args
# 查看上一节点发送的 args.inputData 数据
df = args.inputData
featureColumns = args.featureColumns
labelColumn = args.labelColumn
features = df[featureColumns].values
label = df[labelColumn].values
M = args.M
result = getattr(statsmodels.robust.norms, M)()
" allowed to vary over time.",
)
)
@dc.param(
Bool(
key="mleRegression",
default=True,
help="Whether or not to use estimate the regression coefficients for the"
" exogenous variables as part of maximum likelihood estimation or "
"through the Kalman filter",
)
)
@dc.param(
Int(
key="trendOffset",
default=1,
help="The offset at which to start time trend values.",
)
)
@dc.param(
Int(key="disp", default=5, help="If True, convergence information is printed.")
)
@dc.param(
Int(key="maxiter", default=50, help="The maximum number of function evaluations.")
)
@dc.param(
String(
key="method",
default="lbfgs",
help="The method determines which solver from scipy.optimize is used "
"‘newton’, ‘bfgs’, ‘lbfgs’, ‘powell’, ‘cg’, ‘ncg’, ‘basinhopping’",