Why is MiraWow called "intelligent advertising"?
|Clients:||Business Units or Small Business Enterprises|
At a university late one night, a student is trying to complete his project for the Neural Networks Application Course. During break-time, he watches the news and videos. However, he feels annoyed because the adverts appear suddenly and interrupt his movies. The advertising is also not connected to his desires . At this moment in time, he would prefer to see food adverts (because he is hungry and has no time to cook), or travel adventures (because he is bored).
An idea pops into his mind: “Why are there no apps which can understand you?” An app can listen, watch and appeal to the user’s desires.at that moment. This short anecdote explains the primary aim of MIRAWOW . MIRAWOW is a website/or an app mobile which enables Small Business Enterprises to advertise to their customers intelligently ! MIRAWOW uses a R.A.M.B.O full stack web development procedure and has three artificial brains named Mako, Maeve and Maya .
The benefits of using MiraWow are outlined below:
For users (YOU or customers): MiraWow provides a database of business information and selects the most suitable advertising for YOU. The program considers a number of variables including gender, generation, location and current trends to determine the most appropriate adverts. The advertising will reflect any updated information provided by the business units.
For business units (SMALL BUSINESS ENTERPRISES or Clients): MIRAWOW enables business units to "advertise to their customers intelligently . It enhances the proportion of products that businesses can advertise to their target customers (to the correct customers: “who want to buy”), and reduces the proportion of wrong customers (“feel annoyed”). As a result, their revenue will absolutely increase.
MiraWow means "__L O O K __ W O W__" in Spanish
RAMBO FULL STACK WEB DEVELOPMENT
How is the R.A.M.B.O full stack created?
|Data storage:||Azure/ Firebase/ AWS|
|Back-end||Python/ R studio/ Stata|
|Deployed:||IBM Watson/ Azure/ AWS machine learning|
The R.A.M.B.O (React Azure/AWS Mako and BeyOnt) full stack web was developed by Dr. M. Ha . While creating the MiraWow platform, he recognised that web design is based on several common full stacks such as React-firebase stack, MERM stack, LAMP stack, LEMP stack, MEAN stack, and the Django stack. These full stacks provide a web development roadmap from the front-end (client side) to the back-end (server storage). However, these full stacks DO NOT consider any data science models (such as econometric or machine learning). Therefore, Dr. M. Ha invented a R.A.M.B.O full stack which includes data science models.
Front-end program languages like React/Redux link to data storage server such as Azure blob, Firebase, AWS, Heroku, MongoDB and communicate to users (YOU) and clients (business units). Back-end program languages like Python, and R studio are used to analyze econometric models (such as ordinary lead square/logit regression/GMM regression) and machine learning models (such as random forest, support vector machines, and multi-layer perceptron). All back-end program languages operate in an Anaconda environment. Back-end machine learning models are deployed using an Cloud server(a cloud computing storage service) and connect to the front-end programmes using Nodejs .
Three machine learning models are used to manage the full MiraWow (from front-end to back-end ). These are Mako, Maever and Maya
Team members: MAKO - MAEVE - MAYA
The heart : Understand YOU
MAKO: The heart
This is a hybrid model which combines econometric and machine learning models. First, cluster analysis is used to visualize the data and find any existing clusters. This pre-step uses the Self-organisation map technique. Next, the econometric models (multiple linear regression, logit regression, and GMM regression), are used to investigate significant variables. Insignificant variables are considered ‘noise’ and removed from the original dataset. Last, the remaining variables are trained by machine learning models: random forest, support vector machine, and multilayer perceptron. The final model is optimized and deployed in Cloud storage.
Dr. M Ha’s analysis reveals that his hybrid method increases the classification rate to 1.2%. In particular, the program reduces the Type II error (a customer feeling annoyed) by 50% to 60%. This represents a major reduction: in short, the model significantly decreases the number of unsatisfied customers (by 50% - 60%).
MAEVE: The eye
Maeve is a machine learning model which uses Python. The OpenCV/COCO/or Mobilenet libraries are used to analysis a user’s feeling (YOU). MAKO uses a user’s feeling (sad/happy/angry) to customize the advertising.
MAYA: The ear
This is the last member of the artificial interligent team. This model also uses Python. Tensorflow is used to analyze a customer’s voice and convert it into “key words”. MAKO uses these key words to determine the most appropriate advert or those best suited to the user (YOU).
As a result, three artificial brains MAKO – MAEVE - MAYA work as an intelligent team to enable business units to advertise intelligently
8 " Website modules " are built to generate MiraWow, including:
- 5 Web Platforms : MIRAOC, MIRAFI, MIRAME, MIRASIA, MIRAEU
- 3 Machine Learning : MAKO, MAEVE, MAYA
You can click below images to see these websites
Coming Soon ...FREE DOWNLOAD
The MiraWow app will be available for free download from JULY 2021
MIRA is a business system which includes 7 business lines, including:
- Online : MIRAWOW, MIRACOIN, M.I.J.A.I.
- Offline: MAGRI, MIHOUSE, MICRUISE, MISLAND