With the knowledge that you have gained from this book, you can start to develop your own model that runs on a mobile phone. You will need to identify the problem statement first. There are many use cases where you will not need an ML model; we can't unnecessarily force ML into everything. Consequently, you need to follow a step-by-step approach before you build your own model:
- Identify the problem.
- Plan the effectiveness of your model; decide whether the data could be useful in predicting the output for future, similar cases. For example, collecting the purchase history for people of a similar age, gender, and location will be helpful in predicting a new customer's purchasing preferences. However, the data won't be helpful in predicting the height of a new customer, if that is the data that you are looking for.
- Develop a simple model...