Please enable JavaScript.
Coggle requires JavaScript to display documents.
ML Stages Branch of Artificial Intelligence (AI) - Coggle Diagram
ML Stages
Branch of Artificial Intelligence (AI)
1. Data collection
Initial Step
Different Sources
Databases
Text files
Pictures
Sound files
Web scraping
Data needs to be prepared
2. Data pre-processing
A key step in the process
Involves
Deleting duplicate data
Fixing errors
Managing missing data either by eliminating or filling it in
Adjusting and formatting the data.
Improves data quality
3. Choosing the right model
ML model selection e.g. :
Linear regression
Decision trees
Neural Networks
Depends on:
What kind of data you're dealing with
Your problem
Should be taken into account
when choosing a model to apply
size and type of data
complexity
computational resources
4. Training the model
It means:
Connecting the data to the model
Enabling it to adjust its parameters
Predict output more accurately
Train it with the data that has been prepared
5. Evaluating the model
Important to:
Assess the model's performance before deployment
After model training
It Means:
the model has to be tested on new data that they haven't been able to see during training.
Common metrics to evaluate the performance of a model:
Accuracy in classifying problems
Precision and recall for binary classification problems
Mean error squared with regression problems
6. Hyperparameter tuning
and optimization
May need to adjust
its hyperparameters to make it more efficient after you've evaluated the model
Techniques for hyperparameter tuning
Grid searches
Try different combinations of parameters, and cross-validation
Divide your data into subsets and train your model on each subset, to ensure that it performs well on different data sets
7. Predictions and deployment
Ready to estimate new data:
Once model has been programmed
And, optimized
Done by :
Adding new data to the model
And, using its output for decision-making or other analysis
Deployment of this model involves:
Integration into a Prod environment
where, it is capable of processing real-world data and providing timely information