# Psudo code
def Train(trainFile, modelName):
trainData = load(trainFile)
model = train(trainData)
save(model, "./model/" + modelName)
def Predict(predictFile, modelName):
dataNoApi = load(predictFile)
model = load("./model/" + modelName)
result = predict(dataNoApi)
return dataNoApi + result- File Content
- DataColumns:
["latitude", "longitude", "value", "so2", "no2", "o3", "value2"] - PredictionColumns =
["aqi"] - trainFile contains: DataColumns + PredictColumns
- predictFile contains: DataColumns
- res contains: DataColumns (from PredictFile) + PredictColumns
- DataColumns:
- Model File (
"./model/modelName.pth")- After training, the model can be output into a file
- Loading model
$\rightarrow$ predict$===$ training$\rightarrow$ predict