Question

Problem Description

The first step to solving a problem is defining the problem. For this assignment, you will recognize business problems that may be solved with data mining and you will determine the best data mining technique to solve the problem.

Assignment

For each of the following business problems:

• Pick one of the data mining techniques below to solve the problem

o Classification

o Frequent Pattern Analysis

o Automatic Cluster Detection

• Explain how this technique will solve the problem

• State the business problem as a data mining problem

1. To speed up drive-thru lines, McDonalds wants to predict what drive-thru customers are most likely to order based on the kind of car they drive. You have data on millions of drive-thru orders and you know the type of car that placed each order.

2. You are playing a video game that periodically introduces new characters. When you encounter a character you have not seen before, you must quickly determine if the character is likely to be a friend or a foe. You have lots of data on several hundred characters identified as friend or foe.

3. You work for a very successful high-end company with sophisticated employees who drink wine every time they close a major deal. The company has grown tired of their usual wines and they want you to find new wines they will enjoy. You have data on over 100 wines the company drank in the past and you know whether they liked or disliked each wine.

4. Your company has developed a unique electronics product and they want to identify similar products to help the marketing team develop an effective marketing strategy. You have data on over 1000 electronic devices.

5. The Democratic National Committee wants to analyze voters’ concerns about President Trump to develop the best one-two punch before the 2020 Presidential Election. For example, if a voter feels strongly about Russian collusion, how likely are they to feel strongly about obstruction of justice? The DNC has collected surveys from almost one million voters asking respondents to list their biggest concerns with President Trump.

Answer 1

1. To speed up drive-thru lines with respect to the car driven by the customers I would choose classification. There are different models of cars produced by a single company. So, classifying cars according to their company and then classifying them again according to their models and then predicting the orders placed by the customers would be the efficient one

2. To identify a character as a friend or a foe we can use Frequent Pattern Analysis. The pattern of the newly added characters can be analysed and then categorize them to identify the characters as foes or friends. There will be similarities in the characters. Friend character will have similarities as well as the foe characters. Using pattern analysis, the characters can be identified.

3.To identify new wines according to the likes of old wines you can use Frequent Pattern Analysis. Using frequent pattern analysis and with the collected data the drinking pattern can be identified for each wine to understand whether the workers might enjoy or not enjoy the wine

4. To identify similar products to help the marketing team develop an effective marketing strategy we use Automatic Cluster Detection. Automatic cluster detection groups the given data as clusters. It groups data that are similar so it will be easy to find similar products