Discourse Type Clustering using POS n-gram Profiles and High-Dimensional Embeddings

Abstract: To cluster textual sequence types (discourse types/modes) in French texts, K-means algorithm with high-dimensional embeddings and fuzzy clustering algorithm were applied on clauses whose POS (part-ofspeech) n-gram profiles were previously extracted. Uni-, bi- and trigrams were used on four 19th century French short stories by Maupassant. For high-dimensional embeddings, power transformations on the chi-squared distances between clauses were explored. Preliminary results show that highdimensional embeddings improve the quality of clustering, contrasting the use of bi and trigrams whose performance is disappointing, possibly because of feature space sparsity.

Intravascular ultrasound underestimates vessel dimensions.

The information gathered with intravascular ultrasound (IVUS) are of great value in endovascular techniques. The aim of this study was to evaluate the reliability of IVUS when measuring vessel dimensions by comparison with an established reference method. The left carotid artery was exposed in 4 pigs (45-55 kg) and two piezoelectric crystals were sutured on the adventitia in the same cross-sectional plane. The distance between them was measured either by IVUS and by sonomicrometers. The mean distance between the two crystals calculated by the sonomicrometer was 4.7+/-0.4 mm (mean systolic distance was 4.9+/-0.2 mm, mean diastolic distance was 4.6+/-0.1 mm). The mean distance between the two targets calculated by IVUS was 4. 5+/-0.2 mm (mean systolic distance was 4.6+/-0.2 mm and mean diastolic 4.4+/-0.2 mm). Regression analysis of the two series of data shows a R(2)=0.9984. IVUS measurements are an average 5% smaller than sonomicrometer measurements (3.6% up to 8.3%) and the difference is statistically significant ( p <0.05). The underestimation of IVUS measurements will affect the accuracy, and probably the long-term outcome, of endovascular procedures.

General practitioners can evaluate the material, social and health dimensions of patient social status

OBJECTIVE: To identify which physician and patient characteristics are associated with physicians' estimation of their patient social status.DESIGN: Cross-sectional ulticentric survey. SETTING: Fourty-seven primary care private offices in Western Switzerland. PARTICIPANTS: Random sample of 2030 patients ≥ 16, who encountered a general practitioner (GP) between September 2010 and February 2011. MAIN MEASURES: PRIMARY OUTCOME: patient social status perceived by GPs, using the MacArthur Scale of Subjective Social Status, ranging from the bottom (0) to the top (10) of the social scale.Secondary outcome: Difference between GP's evaluation and patient's own evaluation of their social status. Potential patient correlates: material and social deprivation using the DiPCare-Q, health status using the EQ-5D, sources of income, and level of education. GP characteristics: opinion regarding patients' deprivation and its influence on health and care. RESULTS: To evaluate patient social status, GPs considered the material, social, and health aspects of deprivation, along with education level, and amount and type of income. GPs declaring a frequent reflexive consideration of their own prejudice towards deprived patients, gave a higher estimation of patients' social status (+1.0, p = 0.002). Choosing a less costly treatment for deprived patients was associated with a lower estimation (-0.7, p = 0.002). GP's evaluation of patient social status was 0.5 point higher than the patient's own estimate (p<0.0001). CONCLUSIONS: GPs can perceive the various dimensions of patient social status, although heterogeneously, according partly to their own characteristics. Compared to patients' own evaluation, GPs overestimate patient social status.

Concise reviews: in vitro-produced pancreas organogenesis models in three dimensions: self-organization from few stem cells or progenitors.

Three-dimensional models of organ biogenesis have recently flourished. They promote a balance between stem/progenitor cell expansion and differentiation without the constraints of flat tissue culture vessels, allowing for autonomous self-organization of cells. Such models allow the formation of miniature organs in a dish and are emerging for the pancreas, starting from embryonic progenitors and adult cells. This review focuses on the currently available systems and how these allow new types of questions to be addressed. We discuss the expected advancements including their potential to study human pancreas development and function as well as to develop diabetes models and therapeutic cells.

PACIC Instrument: disentangling dimensions using published validation models.

OBJECTIVE: To better understand the structure of the Patient Assessment of Chronic Illness Care (PACIC) instrument. More specifically to test all published validation models, using one single data set and appropriate statistical tools. DESIGN: Validation study using data from cross-sectional survey. PARTICIPANTS: A population-based sample of non-institutionalized adults with diabetes residing in Switzerland (canton of Vaud). MAIN OUTCOME MEASURE: French version of the 20-items PACIC instrument (5-point response scale). We conducted validation analyses using confirmatory factor analysis (CFA). The original five-dimension model and other published models were tested with three types of CFA: based on (i) a Pearson estimator of variance-covariance matrix, (ii) a polychoric correlation matrix and (iii) a likelihood estimation with a multinomial distribution for the manifest variables. All models were assessed using loadings and goodness-of-fit measures. RESULTS: The analytical sample included 406 patients. Mean age was 64.4 years and 59% were men. Median of item responses varied between 1 and 4 (range 1-5), and range of missing values was between 5.7 and 12.3%. Strong floor and ceiling effects were present. Even though loadings of the tested models were relatively high, the only model showing acceptable fit was the 11-item single-dimension model. PACIC was associated with the expected variables of the field. CONCLUSIONS: Our results showed that the model considering 11 items in a single dimension exhibited the best fit for our data. A single score, in complement to the consideration of single-item results, might be used instead of the five dimensions usually described.