keras-io/timeseries-anomaly-detection

Timeseries anomaly detection using an Autoencoder

This repo contains the model and the notebook to this Keras example on Timeseries anomaly detection using an Autoencoder.

Full credits to: Pavithra Vijay

Background and Datasets

This script demonstrates how you can use a reconstruction convolutional autoencoder model to detect anomalies in timeseries data. We will use the Numenta Anomaly Benchmark(NAB) dataset. It provides artifical timeseries data containing labeled anomalous periods of behavior. Data are ordered, timestamped, single-valued metrics.

Training hyperparameters

The following hyperparameters were used during training:

• optimizer: {'name': 'Adam', 'learning_rate': 0.001, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
• training_precision: float32

Training Metrics

Epochs	Train Loss	Validation Loss
1	0.011	0.014
2	0.011	0.015
3	0.01	0.012
4	0.01	0.013
5	0.01	0.012
6	0.009	0.014
7	0.009	0.013
8	0.009	0.012
9	0.009	0.012
10	0.009	0.011
11	0.008	0.01
12	0.008	0.011
13	0.008	0.009
14	0.008	0.011
15	0.008	0.009
16	0.008	0.009
17	0.008	0.009
18	0.007	0.01
19	0.007	0.009
20	0.007	0.008
21	0.007	0.009
22	0.007	0.008
23	0.007	0.008
24	0.007	0.007
25	0.007	0.008
26	0.006	0.009
27	0.006	0.008
28	0.006	0.009
29	0.006	0.008

Model Plot

View Model Plot