Over sampling and under sampling are techniques used in data mining and data analytics to modify unequal data classes to create balanced data sets. Over sampling and under sampling are also known as resampling.
These data analysis techniques are often used to be more representative of real world data. For example, data adjustments can be made in order to provide balanced training materials for AI and machine learning algorithms.
One area where over sampling and under sampling techniques are used is for survey research. A survey sample population may be unbalanced in terms of types of participants, which can deter the larger population that the survey is meant to study. By using over or under sampling, the ratios of surveyed characteristics, such as gender, age group and ethnicity, can used to make the weight of the data better representative of the group’s ratios within the greater populations.
Over sampling vs. under sampling
When one class of data is the underrepresented minority class in the data sample, over sampling techniques maybe used to duplicate these results for a more balanced amount of positive results in training. Over sampling is used when the amount of data collected is insufficient. A popular over sampling technique is SMOTE (Synthetic Minority Over-sampling Technique), which creates synthetic samples by randomly sampling the characteristics from occurrences in the minority class.
Conversely, if a class of data is the overrepresented majority class, under sampling may be used to balance it with the minority class. Under sampling is used when the amount of collected data is sufficient. Common methods of under sampling include cluster centroids and Tomek links, both of which target potential overlapping characteristics within the collected data sets to reduce the amount of majority data.
In both over sampling and under sampling, simple data duplication is rarely suggested. Generally, over sampling is preferable as under sampling can result in the loss of important data. Under sampling is suggested when the amount of data collected is larger than ideal and can help data mining tools to stay within the limits of what they can effectively process.