What is the MOST direct approach to solve this problem within 2 days?
Train a custom classifier by using Amazon Comprehend.
Build a recurrent neural network (RNN) in Amazon SageMaker by using Gluon and Apache MXNet.
Train a built-in BlazingText model using Word2Vec mode in Amazon SageMaker.
Use a built-in seq2seq model in Amazon SageMaker.
Explanations:
Amazon Comprehend can quickly build a custom classifier on labeled text data, which is suitable for text classification tasks and can be set up without deep ML expertise, meeting the 2-day timeline.
Building an RNN from scratch in SageMaker using Gluon and MXNet requires extensive configuration and tuning, which is time-consuming and not feasible within a 2-day constraint.
BlazingText with Word2Vec is primarily a word embedding model for generating word vectors rather than a classifier, so additional steps would be required to achieve classification.
Seq2seq models are typically used for sequence translation tasks, such as language translation, and are not ideal for simple classification, making it an inefficient choice for this goal.
I figure that the answer is:
Train a custom classifier by using Amazon Comprehend.