Hello and welcome to EDS short education program.
My name is Adam and I will guide you in this program.
The program contains twelve short videos which are aiming to make you familiar with the concept
of EDS.
Once you walk trough this set of videos you will the following:
What is an EDS system?
What benefits you can get from such a system for your organizations.
What are the steps that needed to implemented in such a system and how many resources it
takes?
We hope you enjoy this set of videos and we encourage you to register in our web site
in order to get access to the full list of education and marketing material related to
water quality events detection.
So, what is an EDS.
EDS stands for Early Detection System.
It is a system which enables you to detect abnormal conditions before the situation becomes
critical.
The most common examples are systems which are used for detecting abnormal credit transactions.
For example, when a credit card has been duplicated or stolen.
Another example are systems which are used to detect abnormal conditions in the operation
of cars or aircrafts.
The main target of EDS is to gain time.
Meaning to inform relevant decision makers that some pending problem may be waiting to
be solved before it becomes a crisis.
Machine learning is a mathematical methodology which enables to learn patterns of behavior
from data.
It is widely used in marketing,and economics.
Scientists refer to two main fields of machine learning.
Supervised learning and un-supervised learning.
The un-supervised learning which is more relevant in this case refers to a situation in which
true events are rare or in some cases have never been seen.
For example, an event of intrusion of contamination to a specific water reservoir may not occur
in the history of a specific site.
Given that how do you expect a computerized system to identify such an event for the first
time?
The answer is given by the following table.
Events can be classified by two attributes.
They can be Bad or Good.
And they can be Common or rare.
most of the time when we collect data about water quality the situation is good and common.
A system which is commonly bad should be replaced.
This leaves us with the rare cases.
Good rare are harmless and in most cases, they happen only in
fairy tales.
The problem are the rare cases which may lead to bad situations.
The aim of an EDS system is to learn to detect the bad rare cases.
The challenge is not to alert when the situation is rare good but only in case of rare bad.
What is needed in order to implement Mindset-Detector EDS?
If you don't have on line sensors there is not much, we can do.
But if you already have on-line sensors installed in your water network, or if you have a SCADA
system which collects all the data to one location, moving forward to the next step
is very simple.
All you have to do is transmit the data to our cloud system.
Mindset-Detector is an Amazon Web Services based system.
We use separate Amazon server for each of our customers.
AWS servers are safe and secure.
Only authorized personal may access the data on the server.
Access is allowed only from a known IP address.
We build the detection models for you.
We tune the models and we backup the system regularly.
The water utility does not have to be concerned about any of the management activities which
make the EDS work.
We take care of that.
The only thing which the user needs to do is to approve events classification.
And what about the price?
No upfront payment.
Monthly fee per actual running monitoring models.
Payment is performed only after you consumed the service.
What are the main benefits that you can get
from an EDS system? 1.
Self-auto learning of the statistical borders.
2.
Recognize anomalous leaps that don't violate the low-high user-limits.
3.
Violating engineering rules.
4.
Detection of long-term patterns before those become
critical and calculate the time to alarm.
5.
Detection of low-quality data (static parameter, unlikely time period
, fixed noise).
6.
Unordinary behavior of the I/O cards.
7.
Detection of abnormal event combination.
8.
Detection of event combinations previously defined as dangerous and calculation their
approximate time of occurrence.
9.
Unreasonable noise distinction in the system.
10.
Detection of unexplained variance between different monitoring stations
What is the next step?
We do not ask you to believe us.
Just contact our technical support.
Send a test data set and we will schedule a demo for you.
Please use the email on screen to create contact with our technical team.
We hope you will enjoy this set of videos and you will be interested to implement our
EDS system.
For more infomation >> Nanoscale Informal Science Education Network | Wikipedia audio article - Duration: 7:45. 
For more infomation >> Cornell Laboratory for Accelerator-based Sciences and Education | Wikipedia audio article - Duration: 4:18. 
Không có nhận xét nào:
Đăng nhận xét