Features

• Get proper pre-processing done automatically

  Well-prepared data is the most important recipe for powerful predictive analysis. DAVinCI LABS automatically performs processing tasks that are essential for Machine Learning, such as missing value handling, variable type selection, and various data transformation.

• Follow the UI that thinks like a data analyst

  Modeling, validating, comparing, reporting.. There are a ton of tasks that a data analyst needs to consider. Just follow DAVinCI LABS's authentic UX and its visualized platform that has been designed from years of research in data analysts' modeling process chain.

• Strengthen your predictive analysis with rich ML algorithm package

  With a few mouse clicks, create powerful prediction models with proven Solidware-modified ML algorithms, no hard-coding necessary.

• Get optimal business rules through automatic clustering

  Create a variety of business rules by discovering many complex patterns hidden within Big Data- All in just one day.

• AI is all about auto-optimization

  Maximize model performance with AI based Algorithm auto-tuning, Ensemble, and Champion algorithm selection.

• DAVinCI understands enterprises' needs

  Even a high-performing model is difficult to operate at enterprises, if modeling results are inexplainable or the model itself cannot be updated with ease. With its original technology, DAVinCI LABS overcomes such obstacles by providing explanation for its predictions, and by automatically extracting JAVA library files for fast and convenient model updates.

Process

UX that navigates like a data analyst.

AI engine proven with excellent results.

Discovering hidden patterns inside diverse datasets requires sophisticated data pre-processing and analytics technology

In DAVinCI LABS, we have bundled years of research and know-how gained from enterprise projects by our world-class machine learning scientists

1. Data Upload
   1. Extract structured data from the internal and external datasources
   2. Assign primary key for collecting and merging data
   3. The target variable along with its respective variables for each PK must have been defined
2. Data Preprocessing
   1. Convert data into numeric format that machine learning can recognize
   2. Smart data preprocessing by automatically detecting missing values
   3. Basic statistical analysis and baseline fitting
3. Training Models
   1. Parameter optimization for machine learning algorithms
   2. Rich library of state-of-the-art algorithms including Deep Learning & GBM
   3. Automatic seleciton of the champion algorithm to maximize performance
   4. Developing outstanding models with just a few clicks
   5. Automatic cluster analysis feature
   6. Automatic grading system by each model
4. Prediction Model & Reporting
   1. Automatic report creation for comparing models
   2. Evaluate individual prediction and statistical analysis in Excel file
   3. List of all fields’ weights with respect to the target
   4. Evaluate the developed model using a new dataset
   5. Download the prediction function as a JAVA library for real-time operation

Keep prediction models up-to-date with DAVinCI LABS's instant retraining feature

1. AI will learn from new incoming data and reflect it in the re-trained model
2. Updated prediction function will be automatically produced for easy system integration & real-time operation

PHASE 1Prediction model generation via Machine LearningTraining models based on past data

Client DB (Big Data)

Finance, commerce, manufacturing, medical, etc.

Model training & analysis

Data pre-processing & pattern recognition

Prediction function (scoring model) generation

Mount prediction function on existing system

PHASE 2Operate & update prediction functionCompute prediction for new data / algorithm retraining

Accumulation of new incoming data

Real-time operation of prediction function

Apply predictions to business strategies

Credit scoring, Churn/fraud/defect detection, Target marketing responsiveness prediction

Algorithm retraining &prediction function update

Use Case

• Credit Scoring
• CRM
• Target Marketing
• Early Warning
• Fraud Detection
• Anomaly Detection

• 
  "Credit card issuance credit scoring system innovation pilot project"
• 
  "Enhanced prediction power for both auto accident severity and frequency"
• 
  "2.7% decrease in annual default rate proved through back-testing"
• 
  "Innovation of credit scoring modeling: over twice the predictive power compared to traditional models"
• 
  "Innovation of credit scoring for mid-risk retail lending"
• 
  "Innovation of credit scoring: ability to predict loan defaults increased by 50%"

Specifications

FAQ

• Q

  What data formats are supported, and how do I upload data?

  A

  Currently, DAVinCI LABS supports separator text or CSV file format, and can analyze data combined into one file. Functionalities such as integrating with a database and extracting data from it will be added in the future.
• Q

  What does 'prediction model' mean?

  A

  A Prediction model is a system that is produced from analyzing big data using machine learning algorithms. It is used for predicting future values of specific objects or fields(variables) based on patterns found during machine learning algorithm training. For example, this means predicting the default rate of a loan customer or estimating the loss rate of an insured person.
• Q

  How much data do you need to develop a good prediction model?

  A

  This depends on what the user wants to predict and the nature of the data, but meaningful analysis is plausible with few hundred data samples that contain target information (specific data that represents what the user is trying to predict).
• Q

  What machine learning algorithms are available?

  A

  Currently, there are seven machine learning algorithms available: Neural Network, Stabilized Deep Net, Gradient Boosting Machine, Ridge Regression, Random Forest, Generalized Linear Model, and Logistic Regression. More will be added going forward.
• Q

  Can you develop a prediction model even if you are a non-expert with no knowledge of machine learning?

  A

  The UI/UX is designed to be intuitive and easy to use for non developers/data analysts/machine learning professionals. Powerful prediction models can be generated with just a few mouse clicks. The solution contract includes free-of-charge training for usage and modeling by default, and continuous on-line and off-line customer support are provided.
• Q

  Can I integrate a prediction model with my internal system and run it in real time?

  A

  An advantage of DAVinCI LABS is that it generates machine learning models that produces prediction values for new data in real time. The user can download prediction models in the form of a Java library or other executable files. Once the download is complete, the user can refer to informational documents to be able to mount the model as a function in his/her existing system. Once mounted, prediction values can be requested to the model in real time. Other questions related to customized system integration can be answered via contact contact@solidware.io
• Q

  With machine learning, staying updated by continuously learning from new data is important. Can this be achieved with DAVinCI LABS?

  A

  The user can re-train models as much as he/she wants, such as daily or monthly, to create an updated prediction model that reflects the latest data patterns.
• Q

  Machine learning has been difficult to use in enterprise business due to lack of explanation for predictions. What about DAVinCI LABS?

  A

  DAVinCI LABS's prediction models are based on complex machine learning algorithms, but it has been designed to generate explicable outputs rather than being a black-box. Through its proprietary algorithms, it helps to explain individual predictions by quantifying the influence that each field(variable) contributed in producing them.
• Q

  Is DAVinCI LABS used only in finance?

  A

  If predictive modeling based on pattern analysis is needed, regardless of the business type, DAVinCI LABS can be put to work immediately. Areas where DAVinCI LABS is applicable are: risk assessment, fraud detection, customer behavior based target marketing, churn prediction, defect detection, complaint prediction, efficiency/performance prediction, disease diagnosis, accident prediction, and more
• Q

  Where can I contact for more questions?

  A

  We will gladly provide answers to your questions at contact@solidware.io

Team

Experts in machine learning, data science, and business management from around the globe have gathered here at Solidware.

• 
  Olivier Duchenne, Ph.D
  Co-founder & Tech. advisor
• 
  Sophie Eom
  Co-founder & Business Lead
• 
  Younghan Kim
  Chief Engineer
• 
  Kisung Song
  Chief Designer
• 
  Yury Tsoy, Ph.D
  Chief M.L Scientist
• 
  Christian Lundgren
  Data Engineer
• 
  Minkyu Kim
  M.L Scientist
• 
  Eunseop Jung
  Data Scientist
• 
  Youngmin Park
  M.L Scientist
• 
  Stephen Kim
  Sr. Business Dev. Manager
• 
  Kevin Eum
  Business Dev. Manager
• 
  Hyunju Choi
  Software Engineer
• 
  Hoyeong Heo
  Software Engineer
• 
  Yonghwan Cho
  Data Analyst
• 
  Jiwon Eom
  Business Support. Manager

Contact