Reference no: EM133553471 , Length: word count:1200
TOPIC - STATICAL INFERENCE AND MACHINE LEARNING
Introduction
In this part of the assessment work, you will continue your analysis by focusing on the statistical inference, deciding whether you will perform estimation, and or hypothesis testing, and finally applying appropriate inferential analysis and machine learning techniques.
Details
By now you would have completed preliminary graphical and non-graphical exploratory data analysis of the dataset you have selected. In this assessment activity, you will continue to perform deeper analysis on your data. You will look for more in-depth insights in the data using statistical inference and machine learning techniques you have learned in Topic 3 and 4A, respectively. The analysis will be progressive using a wide range of techniques from simple to more advanced, and each step should help you gain better understanding of the data and more meaningful outcomes for the next stages. An important decision that you will make is to whether to involve statistical interference for hypothesis testing, predicting a target variable, or both. The assumption is that you are familiar with some of machine learning techniques and able to apply them at this stage, if appropriate. However, you will be required to use more advanced machine techniques in the next assessment.
Requirements
At this stage, you will extend your data analytics work by using different techniques as follows. The following guidelines provide clues for more effective analysis.
o Review the insights obtained from the exploratory data analysis in the previous stage. These insights will be useful for informing and setting up your inferential analysis.
o Review the original business problem or issue you were looking a solution for. From the EDA analysis, evaluate if the data (or variables) are appropriate for its intended purpose i.e., solving the business problem. At this stage, you may need to evaluate issues such as additional data that may be required/available for analysis, refining the original business problem or questions etc.
o Examine the effectiveness of calculating correlation coefficients (or correlation matrix) for your data. What variables will you include? Recall that Pearson correlation coefficient is calculated for numerical and ordinal variables but not categorical variables. There are some other techniques such as Spearman' Rank-Order Correlation. What does the result indicate?
o Investigate the suitability of using boxplots, scatterplots, distribution charts and histograms for your data. What insights are you able to derive?
o Apply any other techniques which may be suitable and help refine your questions. Try to make your questions more specific ensuring that they lead to valuable information and insights.
o Investigate the applicability of applying inferential statistics either for prediction or hypothesis testing. Discuss and justify your decision.
o Based on the outcome of previous stage, perform one or both of the following:
o Develop and test at least one hypothesis.
o Discuss the applicability of machine learning techniques including regression analysis and build a model and evaluate it.
o With detail and critically discuss your analysis including the statistical models developed, hypothesis tests, interpretation of the results, etc. supported by evidence (e.g., graphs).
Deliverables
Prepare a report that includes the following:
o A brief description of your business scenario, and outcome of previous stage (100-200 words)
o With detail and critically discuss your work including the use of different analytical techniques, statistical models developed, hypothesis tests, interpretation of the test results, etc. in 800-1000 words.
o Attach insightful graphs, tables etc. that you have developed to support your discussion.
o Ensure that your analysis and report are informed by relevant literature and industry best practices, where appropriate.
Expected total report word count: 900-1200 words.