Tensorflow Sequences Time Series And Prediction

In this fourth course, you will learn how to build time series models in TensorFlow. You’ll first implement best practices to prepare time series data. You’ll also explore how RNNs and 1D ConvNets can be used for prediction. Finally, you’ll apply everything you’ve learned throughout the Specialization to build a sunspot prediction model using real-world data

Where There Are Look At Time Series

Soft Prices

Weathers Forecast

Historical Trend

Time Series

It is order of sequence of values that are usuelly equaly space of time.

Univariate Time Series

Multivariate Time Series

Mouvement Of The Body

Directional Of The Car Of The Univariate

What Kind Of Thing Can We Do With Time Series

1. Prediction Of The Forecasting Base of the data.

2. Projection back on to the time, this proces is call Imputation

3. How is your data base on others data

4. Time Series can be use to detected anormaly: for exapmple on web site log we can to see what is the problem.

1. Other option is to analyze time series to spot of find a pattern what generate the serie itself.

Fixed Partitioning

Roll-Forward Partitioning

Metrics

Naive Forecast MAE

Moving Average

Differencing

Moving Average On Differenced Time Series

Restoring The Trend And Seasonality

Smooting Both Past And Present Values

Preparing features and labels

Window Method

Drop_remainder Parameter

Numpy Method

Split the data into features and label

Shuffle Method

Batch Method

Feeding windowed dataset into neural network

Single layer neural network

Split Data

Simple Linear Regression

Compile The Model

Weight of Model

Predict

Plot A forecast

Measure MSE

Deep neural network training, tuning and prediction

MSE

Callbacks

Plot loss

Recurrent Layer

Sequence to vector

In Practise

Two Layers

Lambda Layers

Full RNN Code

Useful Link

• https://en.wikipedia.org/wiki/Huber_loss

LSTM

Sell State

Bi Directional Cell State

Example of code LSTM

Convolution

Set Windows

Bi-directional LSTM

More About Bi-direction LSTM

More About Optimization Algorithm

• https://www.youtube.com/watch?v=4qJaSmvhxi8

---

Real World Data

How To Read Data

Convert List To Series

Split Data

Window Dataset

Simple DNN

Prediction

Improve Accuracy

Time Series Prediction RNN

Improving RNN Layers

---

What next?

• TensorFlow: Data and Deployment Specialization
• TensorFlow: Advanced Techniques Specialization