User:Shg7D1/Choose an Article

Article Selection
Please list articles that you're considering for your Wikipedia assignment below. Begin to critique these articles and find relevant sources.

Option 1

 * Empirical Distribution Function
 * Clarification of notation would be helpful (what is n? What is the q-th quantile? What does the notation $$x_{(1)}, ..., x_{(n)}$$ refer to? What are all these results being referring to under the asymptotic properties subsection?)
 * Clearer wordings (sometimes later text will refer to "the definitions given above" when we've had 4 or 5 paragraphs of formulas)
 * Statistical implementation does not include the actual functions in the statistical modeling software (in most examples).
 * Sources
 * Sources
 * Sources

Option 2

 * Minimax
 * - Doesn't acknowledge that minimax and maximin can be used to refer to the same problem as well
 * - Examples that use trees, instead of just grids?
 * - Distinguish between examples where each player goes simultaneously (without knowledges of the other players moves) versus cases where players take turns
 * - Possible error/misleading definition for zerosum. Correct definition (I believe): each player's payoff function is the negation of their opponent's such that if player 1's payoff is u_1 and player 2's payoff is u_2, u1 + u2 = 0 (hence zerosum)
 * - Adversarial search?
 * - English difficulty in example "some mix of B1 and B2 instead of either B1 or B2," and "either A1 or A2" instead of "both A1 and A2"
 * - Be clear that minimax equivalent to maximin when dealing with zero-sum games
 * - Citations desperately needed
 * - Combinatorial game theory: tie into zero-sum games?
 * Sources
 * Sources
 * Sources

Option 3

 * Estimation
 * Lead uses unclear wording; maybe better to phrase as
 * Sources
 * Sources
 * Sources

Option 4

 * Supervised Learning
 * "When either type is present, it is better to go with high bias, low variance" why is this? There is no citation for it? If anything, shouldn't this be the other way around because we are trying to avoid modeling the data too closely? High bias could lead to overfitting.
 * Advantage of decision trees is that they handle heterogenous data; disadvantage is that they don't handle continuous variables as well as other algorithms.
 * Give an example? Especially of "features" and classes
 * Add that N is the number of training examples
 * Sources
 * Sources
 * Sources

Option 5

 * Article title
 * Article Evaluation
 * Sources
 * Sources
 * Sources