![](https://acua.qcri.org/blog/wp-content/uploads/2024/07/StockCake-Futuristic-Data-Interaction_1720445493-1024x574.jpg)
Machine learning models have more to offer than what meets the eye. It is often known for uncovering hidden patterns and relationships between variables. The machine learning models provide us with the ability to predict future actions based on the current and past data available. If anything, machine learning is a door that is used to access the skill of making data-driven decision making. To clarify which model is suitable for your research, we should first differentiate the two. The regression model deals with continuous values like height, weight or temperature while the classification deals with distinct values like colours, shoe sizes or exam scores. Dividing the models based on their goals can help clarify what type of analysis they can be used for. For example, regression models aim to explain the y variable against the x variable so it is suitable for analysis that requires understanding the relationship between two variables. Ultimately, classification models aim to classify data into meaningful categories. This would be suitable for analysis that requires classifying products into different categories. Since both models are predictive, whichever one is chosen for the analysis will provide an additional benefit by predicting future actions.
The steps below are to ensure you select the right model:
- Understand your data: by reading the data dictionary you can determine which type of data you are dealing with.
- Select attributes for analysis: narrow down your findings to help focus the analysis on the variables you want to utilize. This will help reduce the data and create more accurate models.
- Choose a model to run: Depending on the attributes selected, you can choose to work with classification or regression. For example, if your attributes are mainly continuous it is recommended to use a regression model.
- Run a performance evaluation to ensure your model is accurate.
Now go ahead and try the models on our exclusive system ACUA!! We offer a platform that helps beginners explore data analytics; try out their latest features now! Link: https://acua.qcri.org/
To gain further knowledge on users, read some research below:
Aldous, K. K., An, J., & Jansen, B. J. (2019). Predicting Audience Engagement Across Social Media Platforms in the News Domain. Lecture Notes in Computer Science, 173–187. https://doi.org/10.1007/978-3-030-34971-4_12
Aldous, K. K., An, J., & Jansen, B. J. (2022). What really matters?: characterising and predicting user engagement of news postings using multiple platforms, sentiments and topics. Behaviour & Information Technology, 1–24. https://doi.org/10.1080/0144929x.2022.2030798