English Version
Machine learning works with specific data types. Each type guides how you clean data, choose models, and build features. Here are the main groups.
1. Numerical Data
Values that represent quantities. You can apply math operations on them.
- Continuous. Any value in a range. Example height and temperature.
- Discrete. Whole numbers. Example counts and items.
2. Categorical Data
Values that represent groups. Models need encoding to use them.
- Nominal. No order. Example colors and cities.
- Ordinal. Has order. Example ratings.
3. Binary Data
Values with two states. Example yes or no. Many classification models use this format.
4. Text Data
Sentences or documents. You convert text into tokens or vectors before training.
5. Image Data
Pixels stored in arrays. Used in vision tasks. Needs preprocessing and augmentation.
6. Audio Data
Waveforms or spectrograms. Used in speech tasks. Needs feature extraction.
7. Time Series Data
Values ordered by time. Used in forecasting and anomaly detection. Needs sequence handling.
8. Tabular Data
Rows and columns. Common in business and analytics. Supports mixed types.
Why This Matters
Each type needs its own preparation steps. Your model choice depends on the data format and structure.
Moroccan Darija Version
Machine learning kaykhdm b data mkhtlfa. Kol type kay7dd tariqa dyal cleaning, models, w features. Hna l types l m3rfin.
1. Numerical Data
Qiyam nssabiya. T9dr tdir 3lihom operations.
- Continuous. Qiyam bin range. Bhal temperature.
- Discrete. A3dad kamla. Bhal counts.
2. Categorical Data
Qiyam dyal groups. Kay7taj encoding bach ykhdm f models.
- Nominal. Bla tartib. Bhal colors.
- Ordinal. Fih tartib. Bhal ratings.
3. Binary Data
Qiyam b jouj states. Bhal yes w no.
4. Text Data
Nass w sentences. Khass conversion l tokens ola vectors.
5. Image Data
Pixels f arrays. Msta3mla f vision. Khass preprocessing.
6. Audio Data
Waveforms ola spectrograms. Msta3mla f speech. Khass extraction.
7. Time Series Data
Qiyam mrrtba b time. Msta3mla f forecasting.
8. Tabular Data
Rows w columns. Msta3mla f analytics. Kayjma3 types mkhtlfin.
3lach hadi mohemma
Kol type kay7taj steps mokhtlfin. Model choice kaybdel 3la hsab data format.