Neurocomputing

Our experimental results are divided into two parts：one：HSC for accuray，which are used to improve the accuracy of the original model ；Two：HSC for model compression，which are used to compress the network

MATLAB .mat file of dynamic simulations of a continuous stirred tank reactor coupled with an external heat exchanger on MATLAB, adapted from the modeling equations in S. Yoon, J.F. MacGregor, Fault diagnosis with multivariate statistical models part I: Using steady state fault signatures, J. Process Control. 11 (2001) 387–400. Data consists of 20 repetitions of 700 samples of 57 predictors (including lagged process variable measurements) and 700 samples of a single response variable for 8 different concept drift models.

1. Due to the large number of pictures, we just selected some of them for display. 2. Original negative sample and Original positive sample are raw data collected from our cooperative unit. We used 100 cervical liquid based cell slides in total, for the sake of simplicity, we have selected a positive and a negative sample for publication, so that you can see the appearance of our raw data which has not been processed. 3. Processed training material are the images which have been processed after binarization, image segmentation and image classification. This folder contains 400 epithelial cells , they are the images of single cells after a series of processes. epithelial cells has been divided into two categories including 200 cancerous epithelial cells and 200 normal epithelial cells, as the names you can see, these are the typical samples we used in the paper.

The code for the proposed MIAM method.

The code of the paper.

This data is the experimental results in our model on AADB database and AVA database.

This is comparison simulation program and results

A demo for SDCH algorithm on NUS-WIDE dataset

Amazon exposes the official datasets (http://jmcauley.ucsd.edu/data/amazon/) which have filtered out users and items with less than 5 reviews and removed a large amount of invalid data. Because of above advantages, these datasets are widely utilized by researchers. We also chose Amazon's dataset for experiments. In our experiments, only users, items, interactions, and category information are utilized. We do the preprocessing in the following two steps: 1. Remove the users whose interactions less than 10 and the items which interactions less than 8 to ensure the effectiveness of each user and item. 2. Select the users with more than 4 sessions, and select up to 90 behavior records for the remaining users. This step guarantees the existence of long- and short-term behavior records and all behavior records occurred within recent three months.

The dataset contains the Chinese and English text data required for the experiment.