For a 13 year old you have made a better video that dives deeper in depth in a way which is completely digestible... and better than most professionals in this space! You've got a bright future ahead of you! I cant wait for your next video.
You are already one of the top 5 ML teachers on youTube due to the clear way in which you explain the concepts. Others know their stuff, but just show you what they know, or sometimes overcomplicate things or go too fast or skip steps. You TEACH better. The animations and your taking each step sequentially and highlighting things help A LOT for us, visual learners. When you´re doing the animations and it gets boring remember they are an essential way to teach faster, and we will thank you later and . You make perfect use of the tools at your disposal to teach. I also liked the absence or memes or music or the image of you talking or anything distracting. Keep it up! Congratulations. You´re a great example of how to use 21st century tools to teach. The more you care about teaching and the less you care about views, the more views you´ll have. Concentrate on continuing providing quality and going in the same path you´re going, whilst working on what you like, because your passion shows. I hope young people like you become teachers in the future because we really need them, particularly in the world of tech where there are so many people that know a lot about a subject but definitely cannot teach their way out of a paper bag and make many learners give up too early..
@@MrEloska Some of my favourites are Sentdex, Data School, Krish Naik, Victor Lavrenko...but if you have some you would recommend please let me know as it's quite hard to find good/clear teachers on these topics
Hi, thank you for your great video! However, I do not understand what is the input data set. I would like to know that from which folder in the database link (CI Machine Learning Repository ) did you obtain your X_data and Y_data?
Thanks for the video, would appreciate some help on how to extract features from new images using python for use in real world application on taking user input and classifying the input into malignant or benign. Or any idea to make this applicable in real world using new cell images.
Why is very important in Binary Classification ML projects to split the data using train_test_split() method from scikit-learn with hyperparameter stratify=True? Explain?
Stratify ensures that training and testing data have representative chunks of each class. For example, if in your dataset, class 1 represents 30% of the data points and class 2 represents 70% of the data points, stratify will ensure that both the training and testing data have these same proportions of class 1 and class 2 (30% and 70% respectively). In general, this isn't a bad practice, especially when your dataset has a lot more of one class than another. In my case, both classes are of similar size. In my code, I randomly shuffle the data before sampling it. With large enough training and testing data, this will almost certainly give you an accurate representation of both classes.
I believe in Classification ML projects the evaluation metrics in order are: Classification Accuracy Score, Confusion Matrix, Classification Report, Area under ROC Curve and Logarithmic Loss. Is this correct?
In this project, each sample had a single number assigned to each attribute, but how about if it was a vector { set of data }? I’m doing a project on fault diagnosis in a small juice plant I built. And each of the faults and the normal behavior have a one minute run in which data are generated ( around 600 data point per run ). Now how do I input this to the neural network ?
At the age of 24, I'm watching this tutorial to have some knowledge about ML whereas I still try to figure out how to go out the light of the fridge after closing the door.
Very impressive that you've learned so much at such a young age. Would you mind telling me your roadmap for how you got to this point (particularly machine learning)? Like where did you start, what resources did you find useful, what order, etc.? I am currently going through Andrew Ng's course and learning the basic ML algorithms but I've never used Keras, TensorFlow, or anything advanced really and would really appreciate seeing how you got to where you are. Also, great video! Thanks!
Like you, I began by learning machine learning (specifically deep learning) concepts. Andrew Ng is a great resource for this. Though I did have some experience with Python (there are many free courses on RU-vid), and this is very helpful when learning ML. I then took a Udemy course on Keras. This was the exact course www.udemy.com/zero-to-deep-learning/ but there are many like it. This video series largely aims to 'get your feet wet', but it is also a great resource (I am not biased at all :))
Thanks! Believe it or not, I made this video all by myself. I began planning the video about three months ago, although I wasn't able to work on it as much as I would have liked because of the demands of school and extracurriculars. Fortunately, I had spare time over spring break, during which I worked on the video a whole lot. It took me a while to learn to effectively use Adobe Premiere Pro, Illustrator, After Effects, and Photoshop (see my older videos to see the evolution of my video style).
Nice video, but please can you help me ,it's very urgent ,also i prepare my file.after load and save with Pickle but i dont know how can use and display this file in website to detect diseases????
The compiled model and scale pickle files need to be deployed in a web server with Python required libraries installed. Then a RESTful API Web Service will be used to consume these file by any needed application. Feel free to read the following blog: “Using C# to call Python RESTful API Web Services with Machine Learning Models” (medium.com/@ernest.bonat/using-c-to-call-python-restful-api-web-services-with-machine-learning-models-6d1af4b7787e). You asked very important question - Thank you!
