Ease of Use
MLS provides a graphical interface that supports drag-and-drop operations, allowing you to easily create workflows for data modeling, analysis, prediction, and visualization.
MLS offers Notebook and supports various open source modeling languages, such as Python and more.
MLS provides abundant machine learning algorithms for you to import and process data, as well as train, evaluate, and export models, covering end-to-end prediction and analysis services.
MLS provides a series of one-stop machine learning applications covering feature engineering, machine learning algorithms, modeling, prediction, and model lifecycle management.
MLS recommends personalized services to customers according to their attributes and behavior features, such as age, gender, work type, marital status, education background, and financial status.
Random Decision Forest Classification (RDFC) and Gradient Boosted Tree Classification (GBTC)
Wealth management recommendation and vehicle price prediction
You can scientifically group customers by data mining and formulate strategies based on customer group characteristics to provide appropriate products and develop target marketing activities, thereby achieving commercial benefits.
Retailers' customer grouping
You can use an automatic network detection system to predict suspicious traffic or faulty devices according to real-time traffic analysis.
PCA-Based Anomaly Detection and Isolation Forest (IF)
Network intrusion detection
You can create a prediction model for devices and provide preventive maintenance suggestions and plans to shorten downtime and reduce the probability of faults, thereby improving efficiency and reducing costs.
Logistic Regression (LR) and Gradient Boosted Tree Regression (GBTR)
Automobile manufacturing and maintenance
EIP access and model lifecycle management
Notebook, helping data scientists with professional analysis
Data visualization, exploring internal data attributes and distribution
NLP, extracting valuable information from texts
Spark 2.1, providing higher performance and better reliability
Preset scenario templates, providing one-click demos
Standard edition, enabling you to create an instance used as computing resources by one click
World Cup prediction model release
Supporting data lake factory service
Supporting automation modeling
MLS provides various nodes that facilitate drag-and-drop creation of modeling workflows.
Data and model visualization functions allow data and machine learning models to be visualized in real time.
MLS provides Notebook that is compatible with third-party development packages and enables data visualization.
Drag-and-drop workflow creation
Simplified and visualized modeling process design allows you to create service analysis and modeling processes without programming.
MLS leverages visualization to achieve fast iteration and convergence so that models can be rapidly deployed in offline production environments.
Data exploration and analysis results can be visualized in various graphs, improving data exploration efficiency.
Trained models and evaluation results are visualized so that they are easy to understand.
The Notebook framework serves as a browser and a unified entry point for data analysis. It can be decoupled from the computing framework of the underlying big data platform and is compatible with open source Python libraries.
Powerful data analysis capabilities
The data analysis process covers data exploration, feature engineering, data modeling, and data visualization.
MLS supports model lifecycle management covering construction, prediction, deployment, and scheduling.
Numerous preset algorithms help you import and process data, as well as train, evaluate, and export models, covering end-to-end prediction and analysis services.
Visualized model lifecycle management
Models of a service are managed according to version. You can manage model and service lifecycles on a visualized management interface.
Industry model standards
Industry-standard PMML files can be imported and exported to be seamlessly integrated with other machine learning software.
Distributed machine learning algorithm library
Large-scale distributed computing increases computing efficiency, allowing models to be trained on larger data sets to increase the accuracy of results.
Automatic data exploration and parameter tuning
Data exploration extracts data types and statistics from raw data. Parameters are automatically tuned for logistic regression and linear regression.