Side Bar/Developpement Bar
Features and usage of side/developement bar
After installing MLBridge, and opening it , a page will pops up and ask you to enter the experiment name, better to choose a name that is close to your data name, for example if your data name is "titanic", then your experiment name could be something like "titanic_exp", this way you can remenber it later easily. a folder will be created under the same name, and when you finish youe experiment all the generated files will be saved in experiment name folder
When you strat mlbridge, the screen will have two sections, side bar and developement bar.
Side Bar
Side bar is the one hilighted in red color, and where you will set the parameters of your experiment:
1 - Specify the ML task you want to perform, could be Classification, Regression,Clustering or Anomaly Detection.
2 - Choose the processing mode, either CPU or GPU
This option is supported only for Classification and Regression.
You can choose GPU option for fast processing, but only if you have a preconfigured one.
In case of Clustering and Anomaly Detection, the app will instead show a list of algorithms to choose from :
Available models for clustering :
‘kmeans’ - K-Means Clustering
‘ap’ - Affinity Propagation
‘meanshift’ - Mean shift Clustering
‘sc’ - Spectral Clustering
‘hclust’ - Agglomerative Clustering
‘dbscan’ - Density-Based Spatial Clustering
‘optics’ - OPTICS Clustering
‘birch’ - Birch Clustering
‘kmodes’ - K-Modes Clustering
Available models for Anomaly Detection :
‘abod’ - Angle-base Outlier Detection
‘cluster’ - Clustering-Based Local Outlier
‘cof’ - Connectivity-Based Outlier Factor
‘histogram’ - Histogram-based Outlier Detection
‘knn’ - k-Nearest Neighbors Detector
‘lof’ - Local Outlier Factor
‘svm’ - One-class SVM detector
‘pca’ - Principal Component Analysis
‘mcd’ - Minimum Covariance Determinant
‘sod’ - Subspace Outlier Detection
‘sos’ - Stochastic Outlier Selection
If the Clustering model you specified is 'ap' or 'meanshift' or 'dbscan' or 'optics', then the number of clusters will be generated automatically.
3 - Import data from csv or excel files,if you select custom data, you can drag and drop your files or browse them. if you select data repository then a list of open source bank data will be available for you for testing.
4 - If the task is Classification or Regression, than you need to select from the list, your appropriate target (Target is the name of the column you want to predict) the process will stop as long as the selectbox is empty, you can choose more than 1 target if you are dealing with multi classification/regression task, in case the the task is Clustering or Anomaly Detection than no target is required.
5 - Enter the session ID number, this is going to be the reference of all your generated file, for example if the session ID = 12 than all names of your generated files will finish by this number (exp : Model_12.pkl )
Note that when ever you want to retrain a new model, you must change the session ID number
6 - Select your activity :
Each of these activities is discussed in details in modules section. it s recommended to go through them one by one, respecting the same order above.
7 - Click save my experiment at the end of the side bar when done experimenting, in this case all your output files will be saved in experiment folder.
Development Bar
Developement bar is the one hilighted above in green color, and where you will execute parameters choosen in Side bar and display the output of your experiment as:
Show your data in a table and describe it.
Generate statistic report and display them.
Setup your data and build your model.
Evaluate your model and visualize the evaluation metrics.
Make prediction and show predicted data.
Export your model
and more which will be explained in Modules section.
Last updated
