R 在GEE中拆分数据集以验证数据点

我在GEE中应用了以下代码，用70%的总训练数据训练分类器

我在导出它时遇到问题。如何正确导出csv，以便在R中使用。我想将预测数据与观察数据导出为csv，用于培训和验证数据集。我导出了csv，变量可以在excel电子表格中看到，我无法应用图1中的代码

    // SVM
var classifier = ee.Classifier.svm().train({
  features: training_subset, 
  classProperty: 'SymbolID', 
  inputProperties: bands
});

// Run the classification 
var classified = image.select(bands).classify(classifier);

print(classified,'classfied')

//create pallet
var Hex =[
  'FF4500',//gums// orange//1
  '14870E',//forest// green//2
  '98FF98',//shrub//mint//3
  'FDF006',//bareground//yellow//4
  'F442D1',//wattle//purple//5
  'FF0000',//pines//red//6
  '08F3E4',//wetland//light blue//7
  'FFFFFF',//urban//white//8
  '0A14F9',//water//darkblue//9
  '00001A',//Shade//black//10
  '808080',//rock//grey//11
  ];

Map.addLayer(classified,{min: 1, max: 11, palette: Hex},
'classification');


// Export classification as raster
Export.image.toDrive({
  image: classified,
  description: 'classification_1',
  scale: 10,
  region: GRD
});


// Generate validation subset
var validation_subset = image.select(bands).sampleRegions({
  collection: Validation_data, 
  properties: ['SymbolID', 'Name'], 
  geometries: true,
  scale: 10
});

print(validation_subset)
// Export predicted versus observed data as 2 csvs for training and validation datasets
// validation 30% not used in training
var validation_30 = classified.select('classification').sampleRegions({
  collection: validation_subset, 
  properties: ['SymolID'], 
  geometries: true,
  scale: 10
});

Export.table.toDrive({
  collection: validation_30,
  description:'classification_1_validation_30',
  fileFormat: 'CSV'
});

我想将以下代码应用于我的csv：

这是导出到我的驱动器的csv