NEXYTE Product Description Page 1 + Geo-format support of geographic coordinate systems with industry standard formats such as Datum WGS84. 4.3 Investigative & Predictive Analytics NEXYTE offers a suite of predictive and investigative machine learning models that accelerate investigation and risk management processes by revealing identities, links, patterns, hotspots, behaviors, and anomalies: + Inferred Relations Analysis. Based on textual analytics and image analytics, NEXYTE generates inferred relations between system entities to reveal and visualize relations derived from text and image sources. + Discovery Analysis. Pattern mapping and clustering of data allow to discern recurring behavior patterns, new developments, and anomalies. + Prediction Analysis. Predictive models can predict or warn of certain behaviors or events based on machine learning, including: Future events such as terrorist attacks, derived from open-source content analysis and messaging group chats Risk profiling of persons for types of crimes such as drug dealing, violence, child abuse, etc. Crime hot spots to prioritize police presence and timing New leads for investigation, based on similarity models + Correlation Analysis. Verification and alignment rules allow to clarify the intelligence picture, providing: Entity resolution - to determine when two profiles are in effect the same person Entity linking - to link a text-derived entity with a system entity Recommendation of entities that may be a match for a suspect 4.3.1 Inferred Relations Image-inferred and text-inferred relations are AI-powered capabilities which detect links between entities within the same picture or document. Image-inferred relations As part of NEXYTE’s image analytics capabilities, image-inferred relations allow the system to detect relations between persons based on their presence in the image file. Intimacy Level Analytics, used as part of the Image-inferred relation capability, allows the categorization of the level of intimacy between the persons in the image using a Machine Learning Model algorithm.