Question

1. Discuss the process of linearization and why you think it is or is not useful.
2. How would you deal with erroneous or “poor” data points when presenting data to fellow scientist?

Answer 1

Linearization simply means to find the linear point of approximation of a function. It is mainly the point where everything starts taking the effect. To perform analysis such as the stability analysis we require linearization so that we can put any model in a linear form.

Note: For the second part of the question, it will be better to provide your opinion from your knowledge of the concept.

Poor data points are a big issue when presenting the data. If the data points are erroneous then there can be issues with high concern later on. So, to resolve this problem one can fix the data at the source use normalization to fix the issues. Sometimes resolving the issues at the first phase is not done then it can be resolved using the ETL phase. Here, the data can go through the Extract, Transform, and Load phases. This can solve the high chances of erroneous data entering the database. And if the redundant or the poor data enters the database then we can use the concept of precision identity to solve it. Precision identity or entity resolution is the most powerful one which helps to solve the redundancy of the data.

For more explanations you can go through the study materials online.