## Overview

A deep learning based customized service for Fast and accurate Time Series data prediction.

Deep Learning and Prediction Process

## Features

Time series data preprocessing

Process raw time series data for more effective deep learning.

Future phenomena forecasting

Predict future outcomes of a system without real experiments.

Optimal deep learning method

Apply the most appropriate state-of-the-art deep learning methods for time series prediction.

Time series prediction tools for easy application

· Provide customers with customized tools for easy application.
· Prediction simply requires executing the tool and entering a design of interest.

## Effects

### Enhance time series prediction accuracy

Supports fast and accurate time series prediction using archived data of various patterns.

### Enhance practical applicability

Customized prediction tool allows easier application to practical problems.

## Applications

### Remaining useful life prediction for Li-ion Batteries

Objective
State of health (SOH) and remaining useful life (RUL) prediction using a deep learning model trained on corresponding archived time series data.

Input Data
Time series data before reaching end of life (EOL)

BruceTS Predictions
· State of Health(SOH)
· Remaining Useful Life(RUL)

Deep Learning Method
TCN(Temporal Convolutional Networks)