As Consultants, Which Of The Following Services Do Forensic Accountants Not Provide?

▸ Unsupervised Learning :

1. For which of the following tasks might K-ways clustering exist a suitable algorithm
   Select all that apply.
   
   
   • Given a gear up of news articles from many different news websites, discover out what are the main topics covered.

     Grand-ways can cluster the articles and then we can inspect them or use other methods to infer what topic each cluster represents
     

   
   
   • Given historical weather records, predict if tomorrow'south weather will be sunny or rainy.
   
   
   • From the user usage patterns on a website, effigy out what different groups of users exist.

     Nosotros tin cluster the users with Chiliad-ways to observe unlike, singled-out groups.

   
   
   • Given many emails, you lot desire to determine if they are Spam or Not-Spam emails.
   
   
   • Given a database of information almost your users, automatically group them into dissimilar market segments.

     You can use G-means to cluster the database entries, and each cluster will correspond to a dissimilar market segment.

   
   
   • Given sales information from a large number of products in a supermarket, effigy out which products tend to form coherent groups (say are frequently purchased together) and thus should exist put on the same shelf.

     If you lot cluster the sales data with G-ways, each cluster should stand for to coherent groups of items.

   
   
   • Given sales data from a large number of products in a supermarket, estimate hereafter sales for each of these products.
   
   

---

2. Suppose nosotros have iii cluster centroids , and .
   Furthermore, we accept a training example . Later a cluster assignment
   step, what volition exist?

---

3. One thousand-means is an iterative algorithm, and two of the following steps are repeatedly carried out in its inner-loop. Which 2?

---

4. Suppose you lot accept an unlabeled dataset . You lot run K-means with 50 different random initializations, and obtain 50 unlike clusterings of the data.

   What is the recommended mode for choosing which i of these 50 clusterings to apply?

---

Check-out our complimentary tutorials on IOT (Cyberspace of Things):

---

5. Which of the post-obit statements are true? Select all that utilise.
   
   
   • On every iteration of Grand-means, the cost function (the baloney function) should either stay the aforementioned or subtract; in item, it should not increase.

     Both the cluster assignment and cluster update steps decrese the cost / distortion role, and so it should never increment after an iteration of K-means.

   
   
   • A good mode to initialize Yard-means is to select K (distinct) examples from the training set and set up the cluster centroids equal to these selected examples.

     This is the recommended method of initialization.

   
   
   • Chiliad-Means will always give the same results regardless of the initialization of the centroids.
   
   
   • Once an instance has been assigned to a particular centroid, it will never be reassigned to another different centroid
   
   
   • For some datasets, the "right" or "correct" value of G (the number of clusters) tin be cryptic, and hard even for a human being good looking carefully at the data to decide.

     In many datasets, different choices of M will give different clusterings which appear quite reasonable. With no labels on the information, nosotros cannot say i is better than the other.

   
   
   • The standard way of initializing K-ways is setting to be equal to a vector of zeros.
   
   
   • If nosotros are worried almost K-means getting stuck in bad local optima, one manner to improve (reduce) this problem is if we attempt using multiple random initializations.

     Since each run of Yard-means is contained, multiple runs can find dissimilar optima, and some should avert bad local optima.

   
   
   • Since K-Ways is an unsupervised learning algorithm, it cannot overfit the information, and thus it is always better to have as big a number of clusters as is computationally feasible.
   
   

---

Click hither to see solutions for all Auto Learning Coursera Assignments.
&
Click hither to see more codes for Raspberry Pi 3 and similar Family unit.
&
Click hither to see more codes for NodeMCU ESP8266 and similar Family unit.
&
Click here to encounter more codes for Arduino Mega (ATMega 2560) and similar Family.

Experience gratis to ask doubts in the comment department. I will try my best to answer information technology.
If you notice this helpful by any mean like, comment and share the mail service.
This is the simplest style to encourage me to keep doing such piece of work.

Thanks & Regards,
- APDaga DumpBox

Source: https://www.apdaga.com/2019/11/coursera-machine-learning-week-8-quiz-unsupervised-learning.html

Posted by: kwonfordlbeatink.blogspot.com

Share this post