C4_W1_Assignment Naive Forecast

Course Q&A TensorFlow Developer Professional Sequences, Time Series and Prediction
1

Naive Forecast

How much values to use for
naive_forecast = series[#YOUR CODE HERE]?

2

Given a time series, say x = [1,2,3,4,5,6,7,8,9,10] (before split), predictions for [6,7,8,9,10] are [5,6,7,8,9].
Naive forecast predicts the value at the previous timestep as the prediction for the current timestep.

1 Like
3

Hi Balaji,
Thanks. What is the reason to build naive forecast?
Best regards, Taras

4

Naive forecast sets a sort of benchmark for the models you would be using
Any predictions cant go worse than this naive forecast

1 Like
5

Thanks Mayank.
What course you can recommend to have better understanding?

6

The course you are doing is one of my recommendations
Other recommendations would be you can refer some youtube channels and also the last course of DLS specialization I think talks about sequential data,LSTM,RNN,etc
Also some kaggle courses might help(These don’t have videos tho)

8

naive_forecast = SERIES[SPLIT_TIME -1:-1]

1 Like