Machine learning models help differentiate between causes of recurrent spontaneous vertigo.

BACKGROUND: Vestibular migraine (VM) and Menière's disease (MD) are two common causes of recurrent spontaneous vertigo. Using history, video-nystagmography and audiovestibular tests, we developed machine learning models to separate these two disorders. METHODS: We recruited patients with VM or MD from a neurology outpatient facility. One hundred features from six "feature subsets": history, acute video-nystagmography and four laboratory tests (video head impulse test, vestibular-evoked myogenic potentials, caloric testing and audiogram) were used. We applied ten machine learning algorithms to develop classification models. Modelling was performed using three "tiers" of data availability to simulate three clinical settings. "Tier 1" used all available data to simulate the neuro-otology clinic, "Tier 2" used only history, audiogram and caloric test data, representing the general neurology clinic, and "Tier 3" used history alone as occurs in primary care. Model performance was evaluated using tenfold cross-validation. RESULTS: Data from 160 patients with VM and 114 with MD were used for model development. All models effectively separated the two disorders for all three tiers, with accuracies of 85.77-97.81%. The best performing algorithms (AdaBoost and Random Forest) yielded accuracies of 97.81% (95% CI 95.24-99.60), 94.53% (91.09-99.52%) and 92.34% (92.28-96.76%) for tiers 1, 2 and 3. The best feature subset combination was history, acute video-nystagmography, video head impulse test and caloric testing, and the best single feature subset was history. CONCLUSIONS: Machine learning models can accurately differentiate between VM and MD and are promising tools to assist diagnosis by medical practitioners with diverse levels of expertise and resources.

Designing a conceptual framework for strategic selection of Bushfire mitigation approaches.

Fires are an important aspect of environmental ecology; however, they are also one of the most widespread destructive forces impacting natural ecosystems as well as property, human health, water and other resources. Urban sprawl is driving the construction of new homes and facilities into fire-vulnerable areas. This growth, combined with a warmer climate, is likely to make the consequences of wildfires more severe. To reduce wildfires and associated risks, a variety of hazard reduction practices are implemented, such as prescribed burning (PB) and mechanical fuel load reduction (MFLR). PB can reduce forest fuel load; however, it has adverse effects on air quality and human health, and should not be applied close to residential areas due to risks of fire escape. On the other hand, MFLR releases less greenhouse gasses and does not impose risks to residential areas. However, it is more expensive to implement. We suggest that environmental, economic and social costs of various mitigation tools should be taken into account when choosing the most appropriate fire mitigation approach and propose a conceptual framework, which can do it. We show that applying GIS methods and life cycle assessment we can produce a more reasonable comparison that can, for example, include the benefits that can be generated by using collected biomass for bioenergy or in timber industries. This framework can assist decision makers to find the optimal combinations of hazard reduction practices for various specific conditions and locations.

Explainable hybrid word representations for sentiment analysis of financial news.

Due to the increasing interest of people in the stock and financial market, the sentiment analysis of news and texts related to the sector is of utmost importance. This helps the potential investors in deciding what company to invest in and what are their long-term benefits. However, it is challenging to analyze the sentiments of texts related to the financial domain, given the enormous amount of information available. The existing approaches are unable to capture complex attributes of language such as word usage, including semantics and syntax throughout the context, and polysemy in the context. Further, these approaches failed to interpret the models' predictability, which is obscure to humans. Models' interpretability to justify the predictions has remained largely unexplored and has become important to engender users' trust in the predictions by providing insight into the model prediction. Accordingly, in this paper, we present an explainable hybrid word representation that first augments the data to address the class imbalance issue and then integrates three embeddings to involve polysemy in context, semantics, and syntax in a context. We then fed our proposed word representation to a convolutional neural network (CNN) with attention to capture the sentiment. The experimental results show that our model outperforms several baselines of both classic classifiers and combinations of various word embedding models in the sentiment analysis of financial news. The experimental results also show that the proposed model outperforms several baselines of word embeddings and contextual embeddings when they are separately fed to a neural network model. Further, we show the explainability of the proposed method by presenting the visualization results to explain the reason for a prediction in the sentiment analysis of financial news.

A systems approach to healthcare: agent-based modeling, community mental health, and population well-being.

PURPOSE: Explore whether agent-based modeling and simulation can help healthcare administrators discover interventions that increase population wellness and quality of care while, simultaneously, decreasing costs. Since important dynamics often lie in the social determinants outside the health facilities that provide services, this study thus models the problem at three levels (individuals, organizations, and society). METHODS: The study explores the utility of translating an existing (prize winning) software for modeling complex societal systems and agent's daily life activities (like a Sim City style of software), into a desired decision support system. A case study tests if the 3 levels of system modeling approach is feasible, valid, and useful. The case study involves an urban population with serious mental health and Philadelphia's Medicaid population (n=527,056), in particular. RESULTS: Section 3 explains the models using data from the case study and thereby establishes feasibility of the approach for modeling a real system. The models were trained and tuned using national epidemiologic datasets and various domain expert inputs. To avoid co-mingling of training and testing data, the simulations were then run and compared (Section 4.1) to an analysis of 250,000 Philadelphia patient hospital admissions for the year 2010 in terms of re-hospitalization rate, number of doctor visits, and days in hospital. Based on the Student t-test, deviations between simulated vs. real world outcomes are not statistically significant. Validity is thus established for the 2008-2010 timeframe. We computed models of various types of interventions that were ineffective as well as 4 categories of interventions (e.g., reduced per-nurse caseload, increased check-ins and stays, etc.) that result in improvement in well-being and cost. CONCLUSIONS: The 3 level approach appears to be useful to help health administrators sort through system complexities to find effective interventions at lower costs.