Auburn University Main Campus False Discovery Rate Discussion Questions

This week we focus on the concept of false discovery in data.  After reviewing the article by Naouma (2019), answer the following questions:

1. What is a false discovery rate?
2. Can a false discovery rate be completely avoided?  Explain.
3. What was the outcome of the results of the use case?

Reply to at least two classmates’ responses by the date indicated in the Course Calendar.

Article for review:

1. Naouma, P. (2019). A comparison of random-field-theory and false-discovery-rate inference results in the analysis of registered one-dimensional biomechanical datasets. PeerJ (San Francisco, CA), 7, e8189–e8189.

post from Anil:

Question 1:

False discovery rate:

The expected balance from type I mistakes is called the false discovery rate (FDR). A type I mistake occurs when the null hypothesis is wrongly rejected, resulting in a false positive. The family-wise error rate is firmly identified with the FDR (FWER). The FWER is the shot reaching somewhere around one erroneous resolution (i.e., one Type I Error). To put it another way, it’s the probability of making any Type I mistake whatsoever. (False Discovery Rate).

The FWER is constrained by the Bonferroni change, which makes preparations for making at least one or multiple false positives. However, this adjustment may be overly rigorous in other fields, resulting in missing findings (Mailman School of Public Health, n.d.). Rather than guarding against making any bogus positive ends, the FDR strategy is an option in contrast to the Bonferroni revision and controls for a low part of fake positives. Thus, factual force is as often as possible upgraded, and type I mistakes are diminished. (Naouma, 2019).

Question 2:

Can a false discovery rate be completely avoided ?

Using a multiple hypothesis correction process like the Bonferroni correction can help reduce the likelihood of any false discovery across all tests, known as the family-wise error rate (FWER). Unfortunately, a well-known disadvantage of the Bonferroni correction and many other FWER control procedures is that they reduce statistical power or the ability to discover genuinely significant phenomena. Furthermore, in interactive data exploration, we must deal with the additional problem that the hypotheses are frequently unknown in advance, making static procedures like the Bonferroni correction ineffective.

Question 3:

Whats the Outcome of the use case:

Data mining’s predictive capability has revolutionized the way company plans are designed. You may now foresee the future by understanding the present. These are a portion of the current business’ information mining use cases and models:

Marketing:

Data mining is being utilized to filter through ever-bigger data sets and further develop market division. It is feasible to foresee buyer conduct by dissecting the relationship between measures, for example, clientage to configuration custom-fitted dedication promoting. In advertising, information mining demonstrates which buyers will probably withdraw from assistance, what intrigues them dependent on their ventures, and what ought to be remembered for a mailing rundown to build reaction rates. Banks use information mining to comprehend market hazards better. It is normal to examine exchanges, card exchanges, buying patterns, and customer financial information in FICO scores and intelligent enemy of misrepresentation frameworks. Data mining also enables banks to understand our online tastes and behaviors better to improve the return on their marketing initiatives, analyze sales channel effectiveness, and manage regulatory compliance duties.

E-Commerce:

Online business organizations utilize data mining to offer strategic pitches and up-sells through their sites. Amazon is perhaps the most notable organization operating information mining strategies to draw in more customers to their eCommerce store. For example, supermarkets use collaborative purchase trends to detect product linkages and determine where they should be placed in the aisles and shelves. Data mining can also decide which deals are most popular with customers or boost sales in the checkout line.

Providers of services. Data Mining can be utilized by service providers such as the mobile phone and utility industries to predict why customers quit their organization. They look at billing information, customer support encounters, and complaints to assign a likelihood score to each consumer and offer incentives.

post from Harshit:

1. What is a false discovery rate?

False discovery data is the ratio of false-positive to total positive results. As per statistics, it is to discover or find out the concept of rate of type 1 errors in null hypothesis testing when performing various comparisons. The proportion of errors committed by false rejecting null hypothesis can be viewed through random variable Q= V/(V+S)- the proportion of the rejected null hypothesis which is rejected in error

1. Can a false discovery rate be avoided entirely? Explain.

Yes, a false discovery rate can be avoided altogether. All positives can be treated as false. If the number of false positives is set to 5%, the total positives maybe 5%. This rate can also be higher because, most of the time, estimators neglect some sources.

1. What was the outcome of the results of the use case?

According to the use case, the method is used to study, find out, clarify the system needs. It comprises interactions between users and systems related to particular systems and particular goals. It also consists of a document that explains the step taken by the user. Each use case has three essential aspects. An actor can be a group of people or a single person. The goal is considered an outcome of the objective which is to be achieved. The system which talks about process and methods followed to reach the outcome or end goal. It also includes their requirements and their expected behaviors.