Day 42 – Forecasting a Time Series and Recurrent Neural Network(RNNs)

One of the most challenging parts is Forecasting a Time Series. If you agree or not, it is pretty hard to predict time-based. If you are working on a project and your task to predict delivery time using already customer-delivered dates. The customer delivery date happens delivery through one truck, and then it is a univariate time series. While in the customer delivery date based on multiple mode truck, ship, airplane, so it is multivariate time series. To predict future dates, which is called forecasting.

To predict future time series, the Recurrent Neural Network(RNNs) can predict the future dates. It can analyze and expect a customer’s delivery when it can happen. Most neural networks are feedforward neural networks; however, Recurrent Neural Network(RNNs) connects backward. Recurrent Neural Network(RNNs) is instrumental in natural language processing applications like speech recognition, speech to text, etc. In one of our blogs, we saw different neural networks and Recurrent Neural Networks (RNNs).

Different types of Recurrent Neural Network(RNNs) and you can see in many ERP systems organizational structure designed like below.

One-to-one
One-to-many
Many-to-one
Many-to-many

We used the most popular activation functions in Recurrent Neural Network(RNNs) in our previous projects.

• Sigmoid

• Relu

• Tahu

We will review and go into more detail into the Recurrent Neural Network(RNNs) in future blogs.

Stay tuned!