- 14
- 4 152
AWA
เข้าร่วมเมื่อ 18 ก.ค. 2016
how to make custom Machine Learning app using rdkit descriptors for predictions
In this tutorial, we demonstrate how to build a predictive model using the RandomForestRegressor from scikit-learn. The app allows you to upload a CSV or Excel file containing SMILES and target columns, visualize the dataset, and choose the target and SMILES columns. You can then explore the feature matrix, select a machine learning model (RandomForest in this case), and even customize hyperparameters using a convenient UI.
Once the model is trained, the app provides insights into the model's performance, including metrics such as R-squared, Mean Absolute Error (MAE), and Mean Squared Error (MSE). Additionally, you can visualize the top 10 important features contributing to the predictions.
But that's not all! Our app also allows you to input a SMILES string, featurize it, and predict the target value. You can export the trained model for later use.
Whether you're a data scientist, chemist, or enthusiast, this app provides an intuitive interface to explore, train, and deploy predictive models for chemical data. Don't forget to like, share, and subscribe for more exciting tutorials on data science and machine learning with Python and Streamlit!"
Once the model is trained, the app provides insights into the model's performance, including metrics such as R-squared, Mean Absolute Error (MAE), and Mean Squared Error (MSE). Additionally, you can visualize the top 10 important features contributing to the predictions.
But that's not all! Our app also allows you to input a SMILES string, featurize it, and predict the target value. You can export the trained model for later use.
Whether you're a data scientist, chemist, or enthusiast, this app provides an intuitive interface to explore, train, and deploy predictive models for chemical data. Don't forget to like, share, and subscribe for more exciting tutorials on data science and machine learning with Python and Streamlit!"
มุมมอง: 70
วีดีโอ
7 | how does lammps works | viscosity wall method | part2
มุมมอง 53ปีที่แล้ว
This video is the second part of the discussion on the viscosity calculation of 2d lj fluid using wall method and can be extended to 3d real world problem. This is done using lammps software. Links to the part1: th-cam.com/video/xxnMZ6xHQ8g/w-d-xo.html
6 | how does lammps works | viscosity wall method | part1
มุมมอง 64ปีที่แล้ว
This video discusses about the viscosity calculation of 2d lj fluid using wall method and can be extended to 3d real world problem. This is done using lammps software.
PSI4 | Part 1 | Basics
มุมมอง 162ปีที่แล้ว
This video describes the usage of psi4. Psi4 is a renowned open-source quantum chemistry software and python interface is also available, designed to facilitate cutting-edge research in quantum chemistry and computational molecular science. This video discuses basic usage of psi4 and explain briefly geometry optimization, energy calculation & interaction energy calculations.
how does lammps work | Part 4 | Calculate viscosity
มุมมอง 751ปีที่แล้ว
Present video discusses how to calculate viscosity using lammps. LAMMPS is a classical molecular dynamics simulation code with a focus on materials modeling. It was designed to run efficiently on parallel computers. In this video series, i will talk about working of lammps. For part 1: th-cam.com/video/6aw6mw-SBY4/w-d-xo.html For part 2: th-cam.com/video/7JwM4n3yYPA/w-d-xo.html For part 3: th-c...
how does lammps work | Part 3 | Viscosity Basics P3
มุมมอง 337ปีที่แล้ว
LAMMPS is a classical molecular dynamics simulation code with a focus on materials modeling. It was designed to run efficiently on parallel computers. In this video series, i will talk about working of lammps. For part 1: th-cam.com/video/6aw6mw-SBY4/w-d-xo.html For part 2: th-cam.com/video/7JwM4n3yYPA/w-d-xo.html
how does lammps work | Part 2 | Basics P2
มุมมอง 75ปีที่แล้ว
LAMMPS is a classical molecular dynamics simulation code with a focus on materials modeling. It was designed to run efficiently on parallel computers. In this video series, i will talk about working of lammps. For part 1: th-cam.com/video/6aw6mw-SBY4/w-d-xo.html
How does lammps work | Part 1| Basic Intro P1
มุมมอง 134ปีที่แล้ว
LAMMPS is a classical molecular dynamics simulation code with a focus on materials modeling. It was designed to run efficiently on parallel computers. In this video series, i will talk about working of lammps.
Introduction to Graph Neural Network GNN from chemistry Point of View | 5
มุมมอง 92ปีที่แล้ว
This video is continuation of the series on Graph Neural networks and discusses Molecules as graph and their application in Machine learning. So stay tuned For part 1: th-cam.com/video/umaw6TRb22s/w-d-xo.html For Part 2: th-cam.com/video/XaW0QF42I0E/w-d-xo.html For part 3: th-cam.com/video/sXbDF2pAK_8/w-d-xo.html For part 4: th-cam.com/video/sXbDF2pAK_8/w-d-xo.html
Introduction to Graph Neural Network GNN from chemistry Point of View | Part 3
มุมมอง 44ปีที่แล้ว
This video is continuation of the series on Graph Neural networks and discusses Molecules as graph and their application in Machine learning. So stay tuned ! For Part 1: th-cam.com/video/umaw6TRb22s/w-d-xo.html For Part 2: th-cam.com/video/XaW0QF42I0E/w-d-xo.html
Introduction to Graph Neural Network GNN from chemistry Point of View | Part 2
มุมมอง 55ปีที่แล้ว
This video discusses graphs and Graph Neural Networks (GNN) from a chemistry point of view. this is the Second video in this series. So keep watching. For First video, Go here: th-cam.com/video/XaW0QF42I0E/w-d-xo.html
Introduction to Graph Neural Network GNN from chemistry Point of View
มุมมอง 72ปีที่แล้ว
This video discusses graphs and Graph Neural Networks (GNN) from a chemistry point of view. this is the first video in this series. So stay tuned
Morgan Fingerprint using rdkit for machine learning application_part2
มุมมอง 5652 ปีที่แล้ว
In this talk, we will see how to generate morgan fingerprint using rdkit from smiles string (In part 1). Link for part 1 th-cam.com/video/vQrzZKHPyms/w-d-xo.html Using these fingerprint we can train supervised learning model, and predict solubility of molecule just given the smiles string.
Morgan Fingerprint using rdkit for machine learning application_part1
มุมมอง 1.7K2 ปีที่แล้ว
In this talk, we will see how to generate morgan fingerprint using rdkit from smiles string. Using these fingerprint we can train supervised learning model, and predict solubility of molecule just given the smiles string.
Can you please make next part of this video. This is highly interesting and you are a good teacher.
Are you a phd student?
How to add more models in this app. Please make video on adding more model
Github Code is here: github.com/AW1176/ML_model_app
Very informative video. Can you please share the code?
Hello, I am using XGboost and I would like to optimize the parameters of the molecular fingerprint. How can I optimize the "Fp_length for a range of (10,,5048) " and "Fp_radius(0,5)"?
Hi. Thanks for watching. To optimize fp length and radius. You need to do grid search, random search or more advanced techniques like Bayesian optimization. U can choose one of these depending on computational resources. Thanks 👍
@@awa8061 Very useful videos! If you don't mind me asking at what point I use grid search on the molecular fingerprint? I am new to this - do you mind sharing an online source to where I can see something similar to what I have to do?
@@ag6228 i thought you asked about optimisation of estimator for which scikit learn is ok. My bad. But as i understand your problem, i believe u need to go through loop. Every time loop runs it takes one variable from your given range e.g. finger print length, and perform XgBoost and give you MAE for regression problem otherwise see evaluation parameters for classification. Then after loop complete you plot MAE for all variables and decide which parameters gave you minimum MAE. This is what i know and it is like a brute force type approach, as I can't think of more smarter approach here. You can share if you know one.
@@ag6228 one hint from computational point, run your loop for 3 or 4 variable and see the response and then explore more. Don't attach full variable space at first. Hope that makes sense
@@awa8061 Ok, now I understand. Thank you very much
Bonjour, merci beaucoup pour vos efforts, je calcule actuellement la viscosité par lammps en utilisant la méthode NEMD mais je ne sais pas comment changer le code en 3 dimensions, s’il vous plaît aidez-moi dans ce sujet
Dimension 3
Oui en 3 d
now the sound is much better & so is the content
Please try to minimize echo from your video. Otherwise, it is good
loved it
❤❤
👍
Nyc....
Error Correction: @ 1:25 It is said that "Adjacency matrix is permutation invariant" which is NOT correct. Correction: "Adjacency matrix is not permutation invariant"
👍👍👍
Thats very good Keep it up <3
Nyc.....
💕
Vvvv informative
Good work Keep it up 💕
Thanks 😊
Hi, If I wanted to create a representative fingerprint from some bits, is there any way? I have some bits of Morgan fingerprint which have the highest impacts on my model and I want to create a fingerprint from them.
Hi faride abd Generated fingerprint are specific for a particular structure distinguished from other molecule. But if you are interested in some specific part of molecule. Lets say carboxyl group. Then your model will be highly parameterized for that part unless you have a very good reason to do that. If i understand you correctly you want to focus on particular functional part. If that's right then there are way. U can even visualize the bits for the molecule. Go ahead and do that. If you can't find then let me. I will b happy to discuss more
Good job You should make a guide on python too