AI/ML Studio
AI Studio provides ML as-a-service capabilities to the Digital Enabler platform, allowing Data Scientists, Data Engineers and Developer to add artificial intellingence to their own applications and analytics.
AI STUDIO API
The AI Studio home page provides a Swagger UI with the documentations of the different APIs. These are grouped into the following categories:
- Matrix Operations
- Linear Models
- Clustering Algorithms
- Graph Algorithms
- Timeseries Analysis
The APIs available for each of the up-mentioned categories will be described in the following sections.
Matrix Operations
Machine Learning and Deep Learning strongly rely on matrix operations and linear algebra. Users can develop their own algorithms using the low-level matrix operations listed below:
Binary operations
- Add
- Dot multiplication
- Elements multiplication
- Matrix power
- Subctract
Unary operations
- Determinant
- Eigenvalues
- Eigenvectors
- Rank
- Transpose
Generators
- Identity Matrix
- Ones matrix
- Random matrix
- Zeros matrix
Linear Models
Linear models are ones of the most used models nowdays. This module provide APIs to fit linear models on provided data and to make predictions using them.
Linear regression
- Fit
- Predict
Clustering Algorithms
Clustering algorithms are unsupervised learning algorithms which aim is to spot unseeing patterns on provided data. The following algorithms are available in AI Studio:
- KMeans
- DBSCAN
- Meanshift
Graph Algorithms
Graphs model social connections and connected data. The following apis allow end users to work on graphs easily:
Shortest Path
- Dijkstra
- Floyd-Warshall
- Johnsons
- Bellman-Ford
- Auto
Additional Operations
- Depth First Order
- Breadth First Order
- Minimum Spanning tree
Timeseries Analysis
Predicting trends is an hot-topic in different domains. There are multiple ways to model temporal data and to make predictions based on the past. The following Algorithms are natively included in AI Studio:
- Forecast (ARIMA/SARIMA)