Feature Selection for Machine Learning (Udemy.com)

Curso bueno para aprender las diferentes metodologías de selección de variables para la creación de modelos de Machine Learning. Aunque se avisa en los prerrequisitos que es necesario tener un conocimiento previo en la construcción de los principales tipos de modelo, sería recomendable incluir una breve sección de repaso para ayudar a las personas que no realizamos estas actividades en el día a día. No obstante, la explicación de todo el contenido del curso muy clara, directa y concisa. ¡Gracias!

The course provides a comprehensive list of the feature selection methods used in machine learning. The lectures include detailed explanations, and the notebooks contain easy-to-follow codes with lots of insightful comments. However, the course is a bit overwhelming. It would be better to present fewer methods and focus on those that work best in practice.

The course was very well structured and explained. Also the provided materials are helpful (presentations as well as jupyter notebooks). The most important information for me was the usage-assessment in the real-world by the tutor at the end of the chapter. I saw also that the tutor is very active when the students asking questions...that is helpful in addition. Thanks a lot!

I really liked this course. It is brilliantly presented, includes top quality Jupyter Notebooks as reference, and it covers the breadth of feature selection methods. The course builds logically, starting with filter methods, then wrapper, embedded, and finally hybrid methods - all in one place. I haven't used hybrid methods or feature engine, so I'll definitely try them on my next project.

I never knew there was so much science underlying feature selection until I completed Soledad's feature selection course. Filtering methods are more straightforward - eg filter out constant features, keeping 1 out of a few correlated features. The wrapper, embedded and hybrid methods are more challenging and take time to understand.

The material was useful. My suggestion to improve the course is to include a discussion on practical aspects of feature selection. For example, if X1 and X2 are correlated features that could be used to predict a response Y, but X2 is more costly to obtain, X2 is more invasive, or X2 is a binary score while X1 is continuous, then X1 seems like the better choice.

Presenta de forma ordenada y comprensible todos los métodos que se pueden utilizar para seleccionar las variables, con sus pros y contras, de forma práctica y reproducible. Merece la pena hacer este curso y tenerlo para consultarlo. Se agradecería mucho que también lo tuviéramos en español. Gracias por el curso Soledad.

This course is just beyond my expectations. Not only this course,but all the courses by the instructor. She's has got so much practical experince,that she guides you if you've any problem to apply the techniques. She shares her experince. You'll get every response of your question in the Q&A forum, and she doesn't only explains it but rather elaborates everything. I'm putting this review after completing the course, and trust me if you try to learn Data Science you must have this ideas,and it'll make you different from others. My respect to the instructor for all her efforts.Thank u a lot

This is the fourth course I am completing from Soledad. Like the other it is very practical. It is an efficient way of boosting your own experience with feature selection. The lectures are easy to follow and well explained. Each method has a theoretical lecture accompanied by a practical example typically for classification and regression.

The course shows a wide range of feature selection options while also addressing the pros and cons of each method as well as the conditions. For those who have just started learning machine learning, this can give you a wholistic options on feature selection. Nevertheless, you might have to understand some statistical concepts and the mechanism behind some ML models. For experienced data scientists, you might learn a few alternative methods.

yes,it is useful to me

I have a ton of ML courses and this is content not really found anywhere else on this site. Especially in this much detail.
Totally worth it.