From 8aa0f70ffd6167ea04209e6f480ee451419aeb88 Mon Sep 17 00:00:00 2001 From: yilingo Date: Wed, 8 Feb 2023 16:35:16 +0800 Subject: [PATCH] Update README.md --- README.md | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index d0ad817..f17bc53 100644 --- a/README.md +++ b/README.md @@ -4,5 +4,9 @@ Fast sensitivity analysis based online self-organizing broad learning system (SA This repository contains an implementation Matlab code. At present, we implements a novel fast partial differential-based sensitivity analysis (FPD-SA) approach to make the model more precise and concise. FPD-SA is a general method that can compress any differentiable model. By introducing FPD-SA into BLS, we provide the offline SASO-BLS algorithm for discrete data and extend it to online mode for streaming data. -Here, one can run `SASO_BLS_offline.m` and `SASO_BLS_online.m` to test SASO-BLS on discrete data and streaming data (datasets: TE process). Furthermore, `BLS_FSA_TSA.m` is implemented to compare the performance of BLS, BLS + traditional SA and BLS+FPD-SA. +# Dataset +Tennessee Eastman process (TEP). One can find the dataset [here](https://github.com/YKatser/CPDE/tree/master/TEP_data) or use the processed dataset in '.\DataSet'. Since the start time and end time for each fault are the 161th and 960th samples, there are 480 training samples and 800 test samples for each fault, plus the normal 520 training samples and 800 testing samples, making a total of 4820 training samples and 8000 testing samples. + +# Demo +The script `SASO_BLS_offline.m` is in charged of testing SASO-BLS on discrete data. Samely, `SASO_BLS_online.m` is implemented to test SASO-BLS on streaming data. Furthermore, `BLS_FSA_TSA.m` is implemented to compare the performance of BLS, BLS + traditional SA and BLS+FPD-SA.