MACHINE LEARNING MODEL EVALUATION
Machine learning (ML) is a subset of Artificial Intelligence (AI) that deals with building systems that learn, and improves performance based on the data inputted. ML is a growing field of data science that makes use of statistical methods, and training algorithms to make classifications and accurate predictions. Machine learning makes use of historical data input to make improved future predictions and uncover key insights in data mining projects.
There are two approaches to machine learning: supervised learning and non-supervised learning
Supervised learning: The data scientist serves as a guide in teaching the algorithm in making rightful conclusions. The algorithm is being trained by a dataset that has already been labeled and has a predefined output.
Non-supervised learning: This approach makes use of a more independent method in which a computer learns to identify complex processes and patterns devoid of human close and constant guidance. Unsupervised machine learning trains data that have no labels or specific, defined output.
Model evaluation involves the construction of mathematical models to better understand any data given for evaluation. Model evaluation is a process of using different evaluation metrics in understanding a machine learning model’s performance, as well as its strengths and weaknesses.
In understanding our data, the following metrics can be used in model evaluation:
Precision: measures the proportion of predicted positive instances that were positive.
Precision = True Positive/ True Positive + False Positive
Accuracy: the ratio between the number of correct predictions and the total number of predictions.
Accuracy = Correct Predictions/ All Predictions
AUC (Area under the ROC Curve): is a performance metric for the classification of problems at different thresholds setting. It shows how a model is capable of distinguishing between classes. The higher the AUC, the better the model is at predicting accurately (when predicted 1 is actually 1 and predicted 0 is actually 0).
Confusion matrix (confusion table): is used in visualizing and summarizing the performance of a classification algorithm.
REFERENCES
What is Machine Learning? https://www.oracle.com/ng/artificial-intelligence/machine-learning/what-is-machine-learning/
Various ways to evaluate a machine learning model’s performance https://towardsdatascience.com/various-ways-to-evaluate-a-machine-learning-models-performance-230449055f15
What is Model Evaluation? https://www.dominodatalab.com/data-science-dictionary/model-evaluation#:~:text=Model%20evaluation%20is%20the%20process,a%20role%20in%20model%20monitoring
https://stephenallwright.com/interpret-auc-score/ How to interpret AUC score (simply explained)