AI Intelligent Analysis and Diagnostic Reports

The StarWay Data Insight has built-in powerful AI intelligent analysis capabilities, which can automatically interpret complex statistical model indicators, perform root cause diagnosis of abnormal data, and automatically generate storable analysis reports based on your business scenarios.

1. Preparatory Work: Large Language Model (LLM) Configuration

Before using the AI analysis function, you need to complete the configuration of the large model service provider in the system settings.

• Multi-model Support: The platform supports connecting to various mainstream large models (such as Qwen, DeepSeek, etc.).
• Flexible Switching: You can add configurations as needed and select the most suitable model service provider during analysis.
• Enterprise-level Support: Supports enterprise users to bind exclusive API Keys to ensure data privacy and stability of analysis processing.

2. AI Automatic Cleaning and Root Cause Diagnosis

During data cleaning or model evaluation stages, when faced with obscure statistical indicators (such as T², SPE), AI can help you complete automatic diagnosis with one click and directly provide business-level processing suggestions.

• Automated Workflow: AI will automatically execute the complete analysis process of "analyzing model indicators -> generating multi-dimensional charts -> diagnosing abnormal data".
• Direct Root Cause Analysis: Not only can it mark the Observation ID of abnormal data (such as samples 3394, 2654), AI will also combine data features to infer abnormal reasons from a business perspective. For example, pointing out that a sample may have "serious measurement errors in key chemical indicators" or "numerical unit errors (such as mistakenly recording mg/L as g/L)".
• Clear Operational Recommendations: For each piece of extremely abnormal or original variable space abnormal data, AI will directly provide clear cleaning suggestions such as "must be removed", greatly reducing the threshold for data cleaning.

3. Analysis Report Storage and Historical Traceability

Each AI diagnosis will automatically generate and store a complete structured analysis report, making it easy for you to review, compare, and review at any time.

• Historical Report Management: You can easily switch and manage historically generated "AI Cleaning Diagnosis Reports" in the left panel, recording the report generation time and the underlying large model used.
• Deep Interpretation Combined with Business Scenarios: You can input specific "business scenario descriptions" (such as "wine industry"), and AI will interpret the actual business meaning of the model (such as PLS-DA) in the tone of an industry expert.
• Business-oriented Key Indicators: Translate abstract academic indicators such as Accuracy into easy-to-understand business language like "about 96 out of every 100 wines are correctly classified", making it convenient to report to non-technical teams.

4. Export and Sharing

The cleaning analysis report generated by AI supports one-click export to Markdown format (see the lower left corner of the pop-up window). You can easily integrate it into your internal work reports, PPTs, or other knowledge base systems to maximize the value of data insights.