Outcome_Oriented_Step_Targets.sos
103 lines
4.5 kB
#!/usr/bin/env sos-runner # fileformat=SOS1.0 # Author: Gao Wang and Bo Peng # # Linear-model based prediction # # This script fits linear models # using Lasso and Ridge regression # and summarizes their prediction performance # This script is written in the "outcome-oriented" style, # demonstrating dynamic input, grouped output # and step dependencies. [global] parameter: beta = [3, 1.5, 0, 0, 2, 0, 0, 0] id = [x+1 for x in range(5)] ridge_result = [f'data_{x}.ridge.mse.csv' for x in id] lasso_result = [f'data_{x}.lasso.mse.csv' for x in id] # Simulate sparse data-sets [simulation] depends: R_library("MASS>=7.3") parameter: N = (40, 200) # training and testing samples parameter: rstd = 3 input: for_each = 'id' output: [(f"data_{x}.train.csv", f"data_{x}.test.csv") for x in id], group_by = 2 R: expand = "${ }" set.seed(${_id}) N = sum(c(${paths(N):,})) p = length(c(${paths(beta):,})) X = MASS::mvrnorm(n = N, rep(0, p), 0.5^abs(outer(1:p, 1:p, FUN = "-"))) Y = X %*% c(${paths(beta):,}) + rnorm(N, mean = 0, sd = ${rstd}) Xtrain = X[1:${N[0]},]; Xtest = X[(${N[0]}+1):(${N[0]}+${N[1]}),] Ytrain = Y[1:${N[0]}]; Ytest = Y[(${N[0]}+1):(${N[0]}+${N[1]})] write.table(cbind(Ytrain, Xtrain), ${_output[0]:r}, row.names = F, col.names = F, sep = ',') write.table(cbind(Ytest, Xtest), ${_output[1]:r}, row.names = F, col.names = F, sep = ',') # Ridge regression model implemented in R # Build predictor via cross-validation and make prediction [ridge] parameter: nfolds = 5 depends: sos_step('simulation'), R_library("glmnet>=2.0") input: dynamic(paths([(f"data_{x}.train.csv", f"data_{x}.test.csv") for x in id])), group_by = 2 output: [(f"data_{x}.ridge.predicted.csv", f"data_{x}.ridge.coef.csv") for x in id], group_by = 2 R: expand = "${ }" train = read.csv(${_input[0]:r}, header = F) test = read.csv(${_input[1]:r}, header = F) model = glmnet::cv.glmnet(as.matrix(train[,-1]), train[,1], family = "gaussian", alpha = 0, nfolds = ${nfolds}, intercept = F) betahat = as.vector(coef(model, s = "lambda.min")[-1]) Ypred = predict(model, as.matrix(test[,-1]), s = "lambda.min") write.table(Ypred, ${_output[0]:r}, row.names = F, col.names = F, sep = ',') write.table(betahat, ${_output[1]:r}, row.names = F, col.names = F, sep = ',') # LASSO model implemented in Python # Build predictor via cross-validation and make prediction [lasso] parameter: nfolds = 5 depends: sos_step('simulation'), Py_Module("sklearn>=0.18.1"), Py_Module("numpy>=1.6.1"), Py_Module("scipy>=0.9") input: dynamic(paths([(f"data_{x}.train.csv", f"data_{x}.test.csv") for x in id])), group_by = 2 output: [(f"data_{x}.lasso.predicted.csv", f"data_{x}.lasso.coef.csv") for x in id], group_by = 2 python: expand = "${ }" import numpy as np from sklearn.linear_model import LassoCV train = np.genfromtxt(${_input[0]:r}, delimiter = ",") test = np.genfromtxt(${_input[1]:r}, delimiter = ",") model = LassoCV(cv = ${nfolds}, fit_intercept = False).fit(train[:,1:], train[:,1]) Ypred = model.predict(test[:,1:]) np.savetxt(${_output[0]:r}, Ypred) np.savetxt(${_output[1]:r}, model.coef_) # Evaluate predictors by calculating mean squared error # of prediction vs truth (first line of output) # and of betahat vs truth (2nd line of output) [evaluation_core] parameter: input_files = list parameter: output_files = list input: dynamic(input_files), group_by = 3 output: output_files, group_by = 1 R: expand = "${ }", stderr = False b = c(${paths(beta):,}) Ytruth = as.matrix(read.csv(${_input[0]:r}, header = F)[,-1]) %*% b Ypred = scan(${_input[1]:r}) prediction_mse = mean((Ytruth - Ypred)^2) betahat = scan(${_input[2]:r}) estimation_mse = mean((betahat - b) ^ 2) cat(paste(prediction_mse, estimation_mse), file = ${_output:r}) [evaluation_lasso] depends: sos_step('simulation'), sos_step('lasso') input_files = paths([(f"data_{x}.test.csv", f"data_{x}.lasso.predicted.csv", f"data_{x}.lasso.coef.csv") for x in id]) output: lasso_result sos_run("evaluation_core", input_files = input_files, output_files = lasso_result) [evaluation_ridge] depends: sos_step('simulation'), sos_step('ridge') input_files = paths([(f"data_{x}.test.csv", f"data_{x}.ridge.predicted.csv", f"data_{x}.ridge.coef.csv") for x in id]) output: ridge_result sos_run("evaluation_core", input_files = input_files, output_files = ridge_result) [default] depends: sos_step('evaluation_lasso'), sos_step('evaluation_ridge') output: lasso_result, ridge_result