The project wants to create a model to predict whether a patient should get a biopsy to test whether the patient has cervical cancer. The data they are using are a large number of features including Age, SocioEconomic and Ethnic factors, Sexual Activity, Family History, Use of Hormonal Contraceptives, whether they have had many children and other factors such as smoking, presence of HPV etc.
What I like about the proposal:
It has a clear goal and rich background information. The background information and its
clear goal provides good justification for the project.
It states clearly what features they are using to create the model.
It provides the prediction algorithms they will use in their prediction. It also states clearly how
they will use these algorithms to make the prediction become feasible.
Areas for improvement:
Perceptron or SVM may not work for this prediction model if the datasets are not linear
separable or approximately linear separable. It may be better to try more classification
algorithms to generate the separation boundary.
Although the proposal lists a lot of features it will use in the model, it gives very few details
about the data types and how they will change their data to fit the input of the model.
It could be better to state how the project separates the datasets into training and testing sets
and how it conducts validations to improve the prediction accuracy of their model.
The project wants to create a model to predict whether a patient should get a biopsy to test whether the patient has cervical cancer. The data they are using are a large number of features including Age, SocioEconomic and Ethnic factors, Sexual Activity, Family History, Use of Hormonal Contraceptives, whether they have had many children and other factors such as smoking, presence of HPV etc.
What I like about the proposal:
Areas for improvement: