eldr.ai | ELDR AI Data Structure
How to pass data to ELDR AI via CSV, API or Text
ELDR AI learns from data it is provided. The dynamic Artificial Neural Network under the hood of ELDR AI makes all the links between and within all data it is given,
spotting trends, patterns and relationships that would otherwise be impossible to observe using traditional methods in the same time frame.
The end result is that ELDR AI is able to make accurate predictions & recommendations when presented with new information as well as give fruitful insights.
ELDR AI makes it easy to use many types of alphanumerical and pure numerical data e.g customer, financial, scientific, financial, engineering, process, criminal,
behavioural, legal, literary etc. for Artificial Intelligence and Machine Learning.
You can pass data in via CSV, API(HTTPS/JSON) or Text depending on how you intend to use ELDR AI.
Supervised v Unsupervised Data
In ELDR AI/Machine Learning, there are 2 common types of data: (1) Supervised & (2) Unsupervised.
Both types contain input values e.g. these could be customer data such as age, post code, credit score, purchases etc.
In Supervised Learning we also have a known output(s). e.g. did the customer return for another purchase?
In this instance ELDR AI would work out how the inputs lead to the output(s), therefore in the case of our customer
example ELDR AI could learn to predict what type of customer is going to return based on their age, post code, credit score and purchases etc.
In Unsupervised Learning we don't have the luxury of an output, we simply have a mass of data to learn from. In this instance ELDR AI
would learn how all the customer data is linked by grouping similar customer profiles together.
Let's now have a look at some specific Supervised and Unsupervised examples and ways of providing data to ELDR AI
In all data used in ELDR AI we only allow letters and numbers e.g. no spaces or special characters.
Provide ELDR AI Data via CSV file
Supervised
Click here for an example of Supervised CSV Data.
Please have a look at the downloaded file:
Row 1 is the labels row which are the data fields we want to use for learning:
PostCode
CreditScore
Age
ProductPurchased
PaymentMethod
Churn
(returning/non-returning customer)
Labels must start with a letter
Row 2 is the inputs and outputs row where we tell ELDR AI which fields are inputs and which are outputs:
ip is used for values with a scale e.g. height, weight, age, time (in this case Credit Score and Age).
ipc
is used for non-scale values, categorical data and words e.g. product name, product ID, user ID, colour, a word from a sentence (in this case Post Code, Product Purchased and Payment Method).
op is always a number
and is 0(no/false) or 1(yes/true) for classification cases such as this one, or can be any number e.g. 0.45, 5, 12.1 for purely numerical tasks
s (not shown in the example) allows you to pass in full sentences e.g. a customer review
ip must always be numeric - ipc can be letters, numbers or alphanumerics - s can be letters, number or alphanumerics - op must only be numeric
The remaining rows contain our actual data.
If you haven't already, please have a look at the
Quick Start
guide which will show you how this supervised data is used in ELDR AI
UnSupervised
Click here for an example of Unsupervised CSV Data.
Please have a look at the downloaded file.
You will see the data looks very similar to our supervised data example except that the output column is missing.
If you haven't already, please have a look at the
Working with Unsupervised Data
guide which will show you how this Unsupervised data is used in ELDR AI
API
Data provided by API follows the same structure as CSV above with labels, inputs and outputs, and rows of data.
Data must be provided in JSON format such as in this short example below:
For a larger example please click here (opens in a new window - returns JSON so a JSON viewer is recommended)
Text
Data can be provided as raw text, following the same structure as CSV above with labels, inputs and outputs, and rows of data.
Data must be provided in the format such as in this short example below (please note the "|" at the end of each line apart from the bottom line):
For a larger example please click here (opens in a new window - to view line by line select "view source" in your browser).
That concludes the ELDR AI guide to supplying data for AI/Machine Learning via CSV, API or Text