There was something wrong with fetching your inbox. Please contact support here.
Strategists, sales leaders, and product managers dream of having the ability to predict outcomes, whether it is to plan better, to proactively identify issues and opportunities, or to make better decisions. Being able to make an accurate prediction can lead to competitive advantage, increased sales, improved customer experience, and reduced costs for many businesses.
Haven OnDemand's predictive capabilities enable developers to create and optimize prediction and recommendation models using APIs that involve little more than uploading the dataset and posing a question to be solved by the system, with no prior Data Science experience needed.
More and more businesses are harvesting and warehousing masses of data, and many are struggling to extract timely insights from their big data lakes. In addition to Enterprise data from business information systems, CRM, and Service Management Systems, they are collecting data from internet logs, websites, social media, sensor data from the Internet of Things (IoT), Open Data, and more. Very quickly, these businesses find their Data Science and Business Intelligence data analytics teams are overwhelmed and struggling to meet cross-functional business demands for predictive modelling. This problem has opened the door for HPE Haven OnDemand to solve with prediction and recommendation APIs for everyday developers that do not have a background in data science or data analytics.
The platform leverages a wide set of algorithms including random forest, logistic regression, support vector machines, and Naïve Bayes for analyzing and creating a machine learning model to perform extrapolation. Rather than restricting to particular techniques it tries all suitable techniques to determine the best performing as well as a machine learning optimizer, which automatically chooses the best set of parameters for each algorithm without any user intervention. Once the models are trained, HPE Haven OnDemand automatically chooses the model that is most accurate, while at the same time ensuring that overfitting doesn’t occur.
The API uses machine learning algorithms (based on mathematical formulas and modeling), statistics, and data modeling to analyze the data. It trains multiple algorithms, tests each alternative, and then selects the algorithm that provides the best results for your data. It stores this information as a prediction model.
The API can create two types of prediction model:
The Get Prediction Model Details API returns the details of a specified prediction model. By default, the API returns information about the structure of the training data, as well as information about the performance measures for the trained model (that is, the algorithm chosen, and details of how well the algorithm fits the training data and matches new data).
You can use the performance_measures and structure parameters to select the information that returns in the API results
The Predict API predicts results by using a classification or regression model created by the Train Prediction API. The Predict API generates a dataset in the same structure as the unpredicted data, with an additional
result field that indicates the predicted values for all records. For classification models, the API also returns a
confidence field, which indicates the confidence level of the prediction (between 0 and 1).
Some examples of how this API could be used:
The Recommend API recommends changes to a data set to achieve a required result. After you create a prediction model by using the Train Prediction API, and predict new data by using the Predict API, you might want to achieve a result different from the prediction. You can use the Recommend API to try and change the outcome of the predicted data by specifying which features can be changed, and asking for a specific result. The model returns a changed dataset to comply with the required result.
Some examples of how this API could be used:
The Delete Prediction Model API deletes a prediction model and all the associated details. You can find a list of your existing prediction models by using the List Resources API.
HPE Haven OnDemand provides more than 70 REST APIs for rapid integration in enterprise, mobile, desktop, IoT, augmented reality, virtual reality, and web apps. Reimagine your world and accelerate development with Applied Machine Learning from Haven OnDemand.