- AI Agent وكيل ذكاء اصطناعي
- A system that carries out multi-step tasks inside applications, not just answers.
- Artificial Intelligence الذكاء الاصطناعي
- Systems doing tasks that used to need human intelligence.
- Business Intelligence ذكاء الأعمال
- Turning data into information that supports decisions.
- Classification التصنيف
- Predicting a category, such as fraud or clean.
- Clustering (k-means) التجميع
- Splitting data into similar groups without given answers.
- Confusion Matrix مصفوفة الالتباس
- A table of a model's hits and misses: TP, FP, FN and TN.
- CRISP-DM منهجية CRISP-DM
- The ML project cycle, from business understanding to deployment.
- Dashboard لوحة مؤشرات
- One screen showing the key indicators at a glance.
- Data Cleaning تنظيف البيانات
- Fixing errors, duplicates and missing values before analysis.
- Data Visualization تصوير البيانات
- Presenting data in charts that make it easier to understand.
- Data Warehouse مستودع البيانات
- A database designed for analysis that consolidates many systems.
- DAX لغة معادلات Power BI
- The formula language for measures and calculations in Power BI.
- ETL استخراج وتحويل وتحميل
- Moving data from sources, cleaning it and loading it into a warehouse.
- Feature خاصية
- A known fact about each example that the model uses to predict, such as order count.
- Hallucination الهلوسة
- When a model confidently produces false information.
- k-Nearest Neighbors أقرب الجيران
- Classifies a new case by the majority of its k closest examples.
- KPI مؤشر أداء رئيسي
- A metric tied to an important goal, with a target.
- Linear Regression الانحدار الخطي
- A model predicting a number with a straight line: intercept + slope × feature.
- LLM نموذج لغوي كبير
- A model trained on vast text that generates by predicting the next word.
- Machine Learning تعلّم الآلة
- Systems that learn rules from data instead of having them hand-written.
- MAE متوسط الخطأ المطلق
- The average gap between forecast and actual, ignoring sign.
- Model Drift انحراف النموذج
- A model weakening over time as new data changes.
- Overfitting الحفظ بدل التعلّم
- A model that excels on training data but does poorly on new data.
- Pivot Table جدول محوري
- A tool summarising data by any dimension with drag and drop.
- Precision الضبط
- Of the cases flagged positive, how many really were.
- Prompt برومبت (موجّه)
- The request you write to an AI model.
- Recall الاستدعاء
- Of all positive cases, how many the model caught.
- Star Schema مخطط النجمة
- An analytical design: a fact table in the centre, dimensions around it.
- Supervised Learning التعلّم المُوجَّه
- Learning from examples that include the right answer: classification and regression.
- Target (Label) الهدف
- The column we want the model to predict, such as “did the customer leave?”.
- Test Data بيانات الاختبار
- A held-back part of the data used to measure the model after training.
- Training Data بيانات التدريب
- The part of the data a model learns from.
- Unsupervised Learning التعلّم غير المُوجَّه
- Learning without answers, such as discovering customer groups.