 |
Model development is the next step to dataset creation. It starts with Exploratory Data Analysis EDA where data is analyzed, use different techniques to interpret the collected data to make analysis easier, more reliable and more effective. Feature Engineering plays an important role in machine learning success or failure. Features are important for predictive models. Those attributes which are useful to the model could qualify to be feature. Features should be carefully selected to reduce the curse of dimensionality and that occurs when feature space ˇ°Mathematical Spaceˇ± has many dimensions and data points are spread far away from each other. Therefor depth versus breath analysis will be conducted. The purpose of a feature, other than being an attribute, would be much easier to understand in the context of a problem. A feature is a characteristic that might help when solving the problem. Next, Dataset will be splatted into 2 segments. One is for training and the other for testing and portions will be distributed 80% and 20% respectively. We will be using several Multiclass classification techniques and choose the best performance outcome. Data will be validated using several techniques and tuned to help us in the selection of best fit technique for our model. Once the best performed classification technique has been selected then we will start to test the model and keep on tuning it until we reach to highest possible performance.
|
IEEE/ICACT20210256 Slide.16
[Big Slide]