METHODS: The modified SPEED or M-SPEED is a sequence prediction algorithm, which modified the previous SPEED algorithm by using time duration of appliance's ON-OFF states to decide the next state. M-SPEED discovered periodic episodes of inhabitant behavior, trained it with learned episodes, and made decisions based on the obtained knowledge.
RESULTS: The results showed that M-SPEED achieves 96.8% prediction accuracy, which is better than other time prediction algorithms like PUBS, ALZ with temporal rules and the previous SPEED.
CONCLUSIONS: Since human behavior shows natural temporal patterns, duration times can be used to predict future events more accurately. This inhabitant activity prediction system will certainly improve the smart homes by ensuring safety and better care for elderly and handicapped people.
OBJECTIVES: To analyze similar-purposed registries and their registry-to-SNOMED CT maps, using two national acute coronary syndrome registries as examples, to understand the reasons for mapping similarities and differences as well as their implications.
METHODS: The Malaysian National Cardiovascular Disease - Acute Coronary Syndrome (NCVD-ACS) registry was compared to the Swedish Register of Information and Knowledge about Swedish Heart Intensive Care Admissions (RIKS-HIA). The structures of NCVD-ACS and RIKS-HIA registry forms and their distributions of headings, variables and values were studied. Data items with equivalent meaning (EDIs) were paired and their mappings were categorized into match, mismatch, and non-comparable mappings. Reasons for match, mismatch and non-comparability of each paired EDI were seen as factors that contributed to the similarities and differences between the maps.
RESULTS: The registries and their respective maps share a similar distribution pattern regarding the number of headings, variables and values. The registries shared 101 EDIs, whereof 42 % (42) were mapped to SNOMED CT. 45 % (19) of those SNOMED CT coded EDIs had matching codes. The matching EDIs occurred only in pre-coordinated SNOMED CT expressions. Mismatches occurred due to challenges arising from the mappers themselves, limitations in SNOMED CT, and complexity of the registries. Non-comparable mappings appeared due to the use of other coding systems, unmapped data items, as well as requests for new SNOMED CT concepts.
CONCLUSIONS: To ensure reproducible and reusable maps, the following three actions are recommended: (i) develop a specific mapping guideline for patient registries; (ii) openly share maps; and (iii) establish collaboration between clinical research societies and the SNOMED CT community.
METHODS: A call for papers was announced on the website of Methods of Information in Medicine in April 2016 with submission deadline in September 2016. A peer review process was established to select the papers for the focus theme, managed by two guest editors.
RESULTS: Three papers were selected to be included in the focus theme. Topics range from contributions to patient care through implementation of clinical decision support functionality in clinical registries; analysing similar-purposed acute coronary syndrome registries of two countries and their registry-to-SNOMED CT maps; and data extraction for speciality population registries from electronic health record data rather than manual abstraction.
CONCLUSIONS: The focus theme gives insight into new developments related to disease registration. This applies to technical challenges such as data linkage and data as well as data structure abstraction, but also the utilisation for clinical decision making.
METHODS: A Malaysian clinical embedding, based on Word2Vec model, was developed using 29,895 electronic discharge summaries. The embedding was compared against conventional rule-based and FastText embedding on two tasks: abbreviation detection and abbreviation disambiguation. Machine learning classifiers were applied to assess performance.
RESULTS: The Malaysian clinical word embedding contained 7 million word tokens, 24,352 unique vocabularies, and 100 dimensions. For abbreviation detection, the Decision Tree classifier augmented with the Malaysian clinical embedding showed the best performance (F-score of 0.9519). For abbreviation disambiguation, the classifier with the Malaysian clinical embedding had the best performance for most of the abbreviations (F-score of 0.9903).
CONCLUSION: Despite having a smaller vocabulary and dimension, our local clinical word embedding performed better than the larger nonclinical FastText embedding. Word embedding with simple machine learning algorithms can decipher abbreviations well. It also requires lower computational resources and is suitable for implementation in low-resource settings such as Malaysia. The integration of this model into MyHarmony will improve recognition of clinical terms, thus improving the information generated for monitoring Malaysian health care services and policymaking.