Reply to Discussion

Answer 1

Algorithms are like road maps to do a specific task. So the implementation of a specific method is a piece of code that calculates the Fibonacci sequence terms. Although a simple function to add two integers is in a way an algorithm. The ideal way to work on complex algorithms is to use them as building blocks to solve logical problems more efficiently in the future. You might actually be astonished by the fact that every day individuals read their emails or hear music on the computers, how many complicated algorithms. How quickly an algorithm is one of the most essential features. To solve a problem, an algorithm is frequently easily developed, but it returns to the drawing board, if the method is too sluggish. As the actual rate of an algorithm and the specifics of its implementation vary upon the location of the algorithm, informaticians usually speak about the runtime as opposed to the input size. Algorithms to determine the shortest route between points. Sometimes, however, the fastest computers are too sluggish even with the latest algorithm, with the most powerful heuristics. Situations such as data compression are addressed in another algorithm class (optimizely.com, 2021).

Statistical significance is the likelihood of not being caused by chance by the difference in conversion rates between a certain variation and the baseline. A test result is considered to have statistical significance, or statistically meaningful, if it is probably not produced by chance at a certain level of statistical importance. Statistical importance is a technique to show that a certain statistics are trustworthy mathematically. You will want to be sure that there is a relationship when you decide on the basis of the findings of the tests you perform. In statistical hypothesis testing, statistic significance is most commonly employed.

It’s a controlled learning problem, if it’s an etiquette. It is an uncontrolled learning challenge where unlabeled data are used to find a structure. It is a learning problem for strengthening, as the solution means optimising an objective function through interacting in an environment. It is a regression problem if the output of the model is a number. It is a classification problem if the model output is a class. If the model output is a collection of input groups, this is an issue with clustering. Data is not the final game itself, but raw material for the entire process of analysis (Almaliki, 2019).

References

optimizely.com. (2021). Statistical significance. Retrieved from, https://www.optimizely.com/optimization-glossary/statistical-significance/

Almaliki, Z. (2019). Do you know how to choose the right machine learning algorithm among 7 different types?. Retrieved from, https://towardsdatascience.com/do-you-know-how-to-choose-the-right-machine-learning-algorithm-among-7-different-types-295d0b0c7f60

—————————————————————————————————————————————-

Answer 2

According to the (Nikolaj Tatti, 2012), The usage of different data algorithms provide the explanatory results for the analysis in order to figure out the appropriate mining approach which delivers the most exceptional insights, it is very important to state the differences provided by the different methods gives the different results. There are multiple statistical approaches to measure the performances of the data algorithms. It is more important to note the difference because the results provided by data mining algorithms are complex, also to study weather both the algorithms are providing the same results or how different are they form information perspective. The probability distributions allow the users to find the similarities and dissimilarities between the different algorithms. It depends up on the business use cases, and about the findings like patterns. these algorithms should be flexible enough to show the same performance when exposed to new data, it helps in in studying the comparative analysis in finding the improved accuracy of the algorithms. for example, receiving operative curve/Area under curve (ROC-AUC) is a Key Performance Indicator which is used for Binary classification to compare the performance of the different classifiers by using various threshold values. it is plotted between sensitivity and specificity based on the results provided by classifier like True positives, false positive, true negatives and false negatives. (Shengping Yang PhD, 2017). Generally, analysts and data scientists who are the persons with the proper domain knowledge who determine the choosing of the final appropriate algorithm for the business use case to predict the outcomes.

References

Nikolaj Tatti, J. V. (2012). Comparing apples and oranges: measuring differences between exploratory data mining results.

Shengping Yang PhD, G. B. (2017). The receiver operating characteristic (ROC) curve.