Applying Machine Learning to Intelligence Problems
Comintelli’s CEO Jesper Martell to Speak About Artificial Intelligence and Knowledge Management at KMWorld ConferenceOctober 24, 2017
Use Cases of Machine LearningNovember 7, 2017
All Machine Learning (ML) algorithms require large amounts of training data (“experiences”) in order to learn. Through experiences and feedback from the environment they then start to recognize patterns in the data and can develop a “model” of the world. As new training data comes in, the algorithm is able to improve and refine this model. This works particularly well when it comes to solving three types of intelligence problems, namely: Classification, Recommendation/Prediction and Clustering.
Identifies information as being x, y or z and classifies and/or labels it
2. Recommendation /Prediction
Recommends the best action or the best content. Predicts the probability of future action based on historical data
Finds patterns or similarities in information and segments into clusters. Discovers associations (e.g. people who read certain books)
Solving classification problems involves making observations, such as identifying objects in images and video or recognizing text and audio which helps determine the “who”, “what”, “when” and “where” in large volumes of data. ML can for example identify “trending themes” or subjects in a text, which human analysts can then use to guide their decisions.
ML can also be used for Recommendations/Predictions, such as estimating the likelihood of events and forecasting outcomes. Furthermore, if an analyst wants to predict the likelihood of a competitor launching a new product at a specific point in time, ML can help with that too.
Lastly, ML can be used to segment data into clusters according to association making it possible to automatically detect and extract sentiment on a more accurate level, e.g. is this news positive or negative, and is this a strength or a weakness for us?
Applied to intelligence problems in a Competitive Intelligence environment, this can have extremely positive outcomes. ML can automate many areas of data synthesis and can significantly help in arriving at faster Times to Insight over traditional methods! Using a good tool for this is key to success and will ensure automation, good content, better quality and faster results. Additionally, learning algorithms can be built into the tool design so the functions can essentially be continually improved, automated and more efficient.
This was part five in our new blog series, based on the article “Machine Learning Implications for Intelligence and Insights”, written by Jesper Martell, Comintelli, and Paul Santilli, Hewlett Packard Enterprise.