IM
Intelligent Modeller
IM is a web-based platform designed for statistical modeling and consequently data analysis. It stores all the necessary steps for statistical modeling. It transforms the variables with the necessary functions for modeling. Subsequently, ensures the automatic reporting of the results by applying on the actual data after establishing the algorithmic relationship between the variables. The platform, which has a project/user management module, is available for use without requiring any software programming knowledge. The Descriptive/Visual Data Analysis Module provides an easy way to grasp the results.
Features |
| Current Solutions | IM Solution |
---|
Variable Transformation |
| Manual | Automated
|
Modelling
|
| Manual | Automated |
Reporting |
| Manual | Automated
|
Data Visualization
|
| Yes | Yes |
Deployment
|
| Manual
| Automated
|
Adaptive Learning
|
| No | Yes |
Highly Skilled Data Scientist |
| Necessary
| Optional
|
Time to Production
|
| High
| Low
|
- User-friendly platform without any software or programming knowledge required.
- The model implementation time is reduced from an average of 1 week to minutes.
- Scalable modelling with self-leaning ability.
- Automatic Performance monitoring is in compliance with Basel Standards.
- Reduces the duration of the modelling process 10 times.
- The daily number of predictive models can increase by 4 times.
- Increases performance by up to 62% compared to manual models.
IDR
Intelligent Data Ready
- IDR is a web-based data preparation and automation platform designed for data stewards, admins, data analysts and data scientists.It helps professionals to easily create automated data flows. According to studies data scientists spend around 80% of their time on preparing and managing data for analysis. Also, 57% of data scientists state that cleaning and organizing data is the most troublesome part of their work. IDR helps to easily manage data workflows and reduces the time spent in this process.
Properties | Current Solutions | IDR Solution |
---|
User Interface | Hard to use | Easy to use |
Programming Knowledge
| Necessary | Low
|
Data Preparation Duration | Days | Hours
|
Data Workflow Management | Hard to do
| Easy to do
|
Automation | Hard to do | Easy to do |
Integration to Different Databases
| Hard to do | Easy to do |
- Easy integration to different databases
- Easy to use UI
- No need to be a professional programmer
- Automated data flows
- Users can easily make changes in production environment
- Reduces the duration of the data preparation process 5 times.
IDQ
Intelligent Data Quality
Intelligent Data Quality is a platform that automatically solves the problem of 'Dirty Data' which is the biggest problem in data analysis. According to studies, 97% of the institutions and companies around the world have a 'Dirty Data' problem. IDQ uses statistical calculations and algorithms to determine the anomalies in the data. It sends automated signals to take precautions to prevent the use of dirty data.
Features | Current Solutions | IDQ Solution |
---|
Working Principle | Rule-Based | Self-learning |
Statistical Knowledge
| Necessary | Not Necessary
|
Anomaly - Error Detection Time | Days
| Seconds
|
Cost of Corrupt Data
| High | Zero |
- Controls data flows in any instant
- User independent, self learning algorithms save time and prevent errors in data flows
- Scalable to volume and variability
- Big data compatible
- Human resource need for data quality controls decreases from 750 man-hour to 3 clicks
- Speed of the corrupt data detection decreases from 7 days to 2.5 minutes
- Data size controlled increases from 1 GB to 8 TB
- Corrupt data detection rate increases from almost 0% to 99%
- Installation & Deployment time decreases by 10 times
AFE
Automated Feature Extraction
Automated Feature Extraction enables to create auatomated and supervised extraction of an optimal set of predictive features from transactional datasets without human effort. It helps transforming temporal and relational datasets into feature matrices for modelling.
Features |
| Current Solutions | AFE Solution |
---|
Working Principle |
| Rule-Based | Self-Learning |
User |
| Necessary | Optional |
Cost |
| High | Low |
SQL Output |
| Manual
| Automated
|
Data Dictionary
|
| Manual | Automated |
| | | |
- Rich additional inputs
- Seleciton of correlated features and less data storage
- Fast working time
- Multiple datasource integrations
- Big data compatible
- Increases model performance by up to 62%
- Reduces time spent on feature generating process.