Columbia Data Analytics Boot Camp - Module 19 - Neural Networks and Deep Learning Models - Created a deep-learning neural network to analyze and classify the success of charitable donations.
- How to compare the differences between the tradiational machine learning classification and regression models and the neural network models.
- How to describe the perceptron model and its components.
- How to implement neural network models using TensorFlow.
- How to explain the differences in neural network structures to change algorithm performance.
- How to preprocess and construct datasets for neural network models.
- How to compare the differences between neural network models and deep neural networks.
- How to implement deep neural network mdoels using TensorFlow.
- How to save trained TensorFlow models for later use.
Step 1 Data Pre-Processing :
- Review the names and types of data,
- Drop unnecessary columns,
- Prepare the data for Step 2.
Step 2 Compiling, Training, and Evaluating the Model :
- Compiled the data,
- Trained the model,
- Printed the performance.
Step 3 Optimization of the model :
- Review the performance printed from step 2,
- Make changes,
- Rerun.
| Model Name | Accuracy Performance |
|---|---|
| 2 | 46% |
| 3 | 73% |
| 4 | 72.7% |
The model did not achieve the target model performance of 75%. The 3rd and fourth models were the closest, yet they did not achieve the target result. I had a hard time figuring out which variables were affecting the neural model. I believe with some more time and different reading, I could write a program that achieved the target model performance metric.