Practice Test 4 | Google Cloud Certified Professional Data Engineer | Dumps | Mock Test
You need to create a Food-based application. Users of the app need to say what kind of food they are craving for. The app will determine the dish which the user will like best. Also, the app should recommend the dish based on previous choices, user input, and the popularity of the dish. The dish will then be delivered to the user’s front gate. What machine learning model will you use to train the application for user preferences?
A. Train a model that returns the items that match most characters in the user query. Once the machine learning model is trained it can fetch the data based on user preferences and ratings given by the user.
B. You train a linear model in TensorFlow with a wide set of cross-product feature transformations. This model predicts the probability of consumption of items for each item based on previous experience. So, this model delivers the top item with the highest predicted consumption rate.
C. Create a TensorFlow model with a deep feed-forward neural network. This model will learn lower-dimensional dense representations for every query and item. Thus query will be given a recommendation based on the input given to the app and by matching items to queries that are close to each other.
D. Create a Wide and Deep Learning model in TensorFlow. This model can use cross-feature transformation in the wide model and generalize to similar items using a deep model.
The correct answer is D.
Option A is incorrect. This model will not work as matches could be irrelevant. For eg: people ordering fried chicken will end up getting chicken fried rice.
Option B is incorrect. This model will only recommend users based on their previous orders. The app will not recommend any item which is liked by other users for the same input. For eg: people ordering “fried chicken” do not mind getting “burgers”. This model is totally based on memorization and therefore is not correct.
Option C is incorrect. This model will work on generalization which might work for some users, but for some users who want any targeted product, this model will not be correct. For eg: if some users ask for specific items “iced decaf latte with nonfat milk” might get a “hot latte with whole milk”.
Option D is correct. For these kinds of use cases, a Wide and Deep learning model is required. The TensorFlow model for the same is already created. Please refer below link for more details –
Comments are closed, but trackbacks and pingbacks are open.