The transformational shift of the time!

Let us craft the world where people use intelligent databases to make better data-driven decisions.

…and believe the best way to make these decisions is to enable AI capability within existing database servers without the effort of creating a new one.

Shift BI to AI:

There is an ongoing transformational shift within the modern business world from the “what happened and why” based on historical data analysis to the “what will we predict can happen and how can we make it happen” based on machine learning predictive modelling.

Machine Learning (ML) Lifecycle:

The ML lifecycle can be represented as a process that consists of the data preparation phase, modelling phase, and deployment phase. The diagram below presents all the steps included in each of the stages.

BI Table:

Let us understand a table in a database. A table is an arrangement of information in rows and columns containing cells that make comparing and contrasting information easier.

Let us understand the below SQL query.

SELECT income, debt FROM income_table WHERE income = 80000;

The above query is used to retrieve income and debit columns, from the table income_table, where the row is found with the column income holds 80000.

SQL is the language used to retrieve the information stored in the tabular form in any DB Table. There are various databases that store data in various methods.

Horizontal, Vertical, Key-Value pair, Relational, Graph, and object-oriented databases are the major types of databases.

AI Table:

The first time we would come across something called AI Table. What is AI Table?

In an AI table, the logic is built in such a way when we write a SQL query with conditions where the income that is looked up in the table, is not matching exactly in the table, the debt will be predicted based on the Regression line using Auto ML. The relation between the Independent target variable debt will be predicted based on the Linear regressor being a predictable nature of data type of Debt.

When we apply the SQL query, it will return the value based on the Predictor defined in the AIML Model Layer of Database.

SELECT sqft, price FROM home_rentals_model WHERE sqft = 800;

Result of the Query:

It is very much a smart way for application design and development intelligently return the debt for the given value irrespective of whether the value is present in the table or not.

All this prediction is just happening miraculously using the typical SQL query, which is the beauty of the AI Table. Look at the syntax of the SQL query to get the prediction of debt based on the income vs debt relationship given in the “SELECT” statement.

CREATE PREDICTOR debt_model FROM income_table PREDICT debt;

The above query will return 28010 for the value 90120, though there is no matching record in the table with an income of 90120.

This is because SQL Query runs the model automatically and predicts the value based on the linear regressor.

This kind of smartness can be implemented in all the SQL Servers today in the market using the state of art AutoML packages widely available on the open-source python.

The architecture of the Typical AI Table:

AI Table comprises 3 layers. Data Layer, Prediction Layer and Visualization Layer.

MindsDB:

Introducing the AutoML Prediction Layer for all databases known as “MindsDB”/

MindsDB is a predictive platform that makes databases intelligent and machine learning easy to use. It allows data analysts to build and visualize forecasts in BI dashboards without going through the complexity of ML pipelines, all through SQL. It also helps data scientists to streamline MLOps by providing advanced instruments for in-database machine learning and optimizing ML workflows through a declarative JSON-AI syntax.

MindsDB enables advanced predictive capabilities directly in your Database. This puts sophisticated machine learning techniques into the hands of anyone who knows SQL (data analysts, developers and business intelligence users) without the need for a new tool or significant training.

How to install MindsDB on Local Desktop using Anaconda?

Create the Anaconda Environment and install the mindsdb package.

conda create -n mindsdb

conda activate mindsdb

pip install mindsdb

  1. Let MindsDB connect to your database.
  2. Make MindsDB learn from historical data automatically by training a predictor using a single SQL statement. (if you’d rather configure some of your models manually or bring your own, MindsDB supports that too via declarative JSON-AI syntax).
  3. Make predictions immediately by querying MindsDB virtual AI Tables. There’s no need to deploy models.
  4. Visualize forecasts in your BI dashboards, all through standard SQL. AI Tables behave the same way as normal database tables.

Features of AI Tables:

  • Automatic data pre-processing, feature engineering and encoding.
  • Classification, regression, time-series tasks.
  • Automatic model deployment through virtual AI Tables.
  • Data quality check for potential biases & outliers.
  • Model accuracy scoring and confidence intervals for each prediction.
  • Batch predictions by joining predictors with other tables.
  • Anomaly detection.
  • Model explainability analysis with the graphical user interface.
  • GPU support for faster model training.
  • Tune model internals with declarative JSON-AI syntax.
  • Import your own models.
  • HTTP API available

Thanks for reading the post. Hope you might find it very useful. Thanks for reading. I will come up with some interesting stuff for tomorrow. Please feel free to provide your critics and views.

MindsDB docker to deploy!

Getting Started with MindsDB

Originally published at https://www.linkedin.com.

--

--

--

Lakshminarasimhan, is CEO and Founder of Graspinor. Graspinor is a renowned Brand for AIML consulting. It was founded in 2019 and operating out of Chennai.

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

Indonesia AI Forum Debuts an Open Forum, Points the Way to Increase Awareness and Understanding of…

Talking with Yala, the coolest & mightiest social media bot on Slack.

AI and a renowned dancer performed together — Can AI stimulate the dancer’s creativity?

When a Chatbot is not the Best Start for Customer Service

Eden AI launches platform to unify ML APIs

About the Long Tail…

Separating fiction from reality has become harder than ever with the spread of deep fakes.

Intelligent Process Automation | 2020 Benefits & Use Cases

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Lakshminarasimhan S

Lakshminarasimhan S

Lakshminarasimhan, is CEO and Founder of Graspinor. Graspinor is a renowned Brand for AIML consulting. It was founded in 2019 and operating out of Chennai.

More from Medium

Spark Standalone multinode setup on Windows

Understanding Calculus using NM Dev libraries

Decision algorithm for a selection of correct functions in Anaplan: SUM or LOOKUP Function?

Nature in computer science: Decision trees and Random forests