How are number of features and number of observations related?
Are there any papers/books/ideas about the relationship between the number of features and the number of observations one needs to have to train a “robust” classifier? For example, assume I have 1000 features and 10 observations from two classes as a training set, and 10 other observations as a testing set.
How to find the best predictors for ML algorithms?
According to Sayes et al.¹: “The objectives of feature selection are manifold, the most important ones being: to avoid overfitting and improve model performance, i.e., prediction performance in the case of supervised classification and better cluster detection in the case of clustering,
Can a large number of predictors make a model less interpretive?
Will it make your model less interpretive if you use all the features? Having a large number of predictors are also likely to increase the development and model training time, while at the same time utilizing a large amount of system memory.
Which is the best correlation measure for ML?
Perhaps the best-known correlation measure is Pearson’s correlation that assumes a Gaussian distribution of the data. Correlation coefficients range between -1 (perfect negative correlation) and 1 (perfect positive correlation) with 0 representing no relationship between the variables whatsoever.
Why is sample size important when conducting research?
When conducting research, quality sampling may be characterized by the number and selection of subjects or observations. Obtaining a sample size that is appropriate in both regards is critical for many reasons. Most importantly, a large sample size is more representative of the population, limiting the influence of outliers or extreme observations.
What is the difference between an under and over-sized experiment?
An under-sized experiment exposes the subjects to potentially harmful treatments without advancing knowledge. In an over-sized experiment, an unnecessary number of subjects are exposed to a potentially harmful treatment, or are denied a potentially beneficial one.”
How big should the sample size be for a survey?
(5) Thus, the authors determined that a sample size of 30 respondents would provide a reasonable starting point.