Final answer:
Oversampling using bootstrapping generates duplicate data points from the minority class, while SMOTE creates synthetic data by interpolating between instances. Data visualization techniques like frequency polygons help compare distributions, and visual tools such as word clouds and polls can make data easily understandable.
Step-by-step explanation:
The techniques of oversampling using bootstrapping and SMOTE (Synthetic Minority Over-sampling Technique) are employed to address class imbalance in datasets during the preprocessing step in machine learning. Bootstrapping is a resampling technique that involves randomly sampling with replacement from an existing dataset to create new samples. In the context of oversampling, bootstrapping would enhance the minority class by generating additional data points that are exact copies from the original dataset. In contrast, SMOTE generates synthetic observations by interpolating between existing minority class instances, allowing for a more diverse and richer dataset that does not rely on simple duplication.
For data visualization, there are various techniques like frequency polygons, which help compare distributions across different datasets. Visual tools such as word clouds, surveys, and live polls engage an audience effectively and provide a visual representation of data in an easily understandable manner. In biodiversity studies, for example, visualization can help compare different biodiversity indices and interpret complex datasets in conservation biology.