Overview
During this assessment, you will produce a written report on neural networks using AWS SageMaker.
Learning outcomes
Analyze real world tasks using machine learning methods, in particular describing, choosing, and applying appropriate supervised machine learning methods for descriptive data mining tasks
Develop and deploy machine learning models on AWS
Tune hyperparameters for machine learning models using AWS
Programmatically interact with AWS using Python Jupyter Notebooks
Synthesize and communicate the method and findings to diverse audiences
Format
You will need to submit the following:
A PDF file clearly shows the assignment question, the associated answers, relevant Python outputs, analyses and discussions.
Appendices include Jupyter Notebook/ Python code.
The assignment should not exceed 15-A4 pages. Appendices do not form part of the page limit.
The task cover sheet
Background
Understanding the possibility of readmission of diabetes patients may provide hospitals and doctors insight into effectiveness of ongoing treatments, and potential changes in the treatments. The changes in treatments may potentially save patients’ life. You are asked to investigate if a neural network accurately classifies if a patient with diabetes is likely to be readmitted to the hospital given their current conditions.
Data
The data, ‘’diabetic_data.csv”, contains 10 years (1999-2008) of clinical care at 130 U.S. hospitals and integrated delivery networks. Data description is provided in Table 1, “Impact of HbA1c Measurement on Hospital Readmission Rates: Analysis of 70,000 Clinical Database Patient Records” BioMed Research International, vol. 2014, Article ID 781670, https://www.hindawi.com/journals/bmri/2014/781670/
Assessment Tasks
1. Data preparation
Discuss a subset of relevant predictors used in a NN model
Apply and Discuss any appropriate cleaning or transformations
Apply and discuss the training and testing dataset.
2. Build, train and deploy a neural network
Propose a baseline structure of a multiple layer perceptron (MLP) neural network used for the classification. Justify your choice.
Report and discuss the performance of the proposed model.
Apply techniques such as dropout, early stop, batch normalization to the baseline MLP (in (a)), and investigate their impacts on the performance of MLP.
Discuss limitations of the proposed model for the classification task.
3. Discuss the considerations of using AWS SageMaker. At least include discussion regarding:
Provide evidence of successfully training, deploying and creating endpoints of all models on AWS Sagemaker.
Notebook instance type
Cost and computation time
For Complete solution of this sample assignment mail us on contact@codersarts.com
Also, If you are looking for any kind of coding help in programming languages, Contact us for instant solution
Comments