What is I C A ?
In the framework of multi-domain Text-to-Speech synthesis it is essential to (i) design a hierarchically structured database for allowing several domains in the same speech corpus and (ii) include a text classification module that, at run time, assigns the input sentences to a domain or set of domains from the database. In this paper, we present a hierarchical text classifier based on Independent Component Analysis (ICA), which is capable of (i) organizing the contents of the corpus in a hierarchical manner and (ii) classifying the texts to be synthesized according to the learned structure. The document organization and classification performance of our ICA-based hierarchical classifier are evaluated in several encouraging experiments conducted on a journalistic-style text corpus for speech synthesis in Catalan.
What is the strong point of I C A ?
For complex-valued multidimensional signals, conventional decorrelation methods do not completely specify the covariance structure of the whitened measurements. In recent work [1,2], the concept of strong-uncorrelation and its importance for complex-valued independent component analysis has been identified. Few algorithms for estimating the strong-uncorrelating transform currently exist. This paper presents two novel algorithms for estimating and computing the strong uncorrelating transform. The first algorithm uses estimated covariance and pseudo-covariance matrices, and the second algorithm estimates the strong uncorrelating transform directly from measurements. An analysis shows that the only stable stationary point of both algorithms produces the strong uncorrelating transform when the circularity coefficients of the sources are distinct and positive. Simulations show the efficacy of the approach in a source clustering task for wireless communications.
What is the weak point of I C A ?
The main reason to determine estrogen (ER) and progesterone receptors (PR) in breast cancer is their predictive value for the response to endocrine therapy. In addition, ER and PR are often used as prognostic indicators. Enzyme immunoassay (EIA) and immunocytochemical assay (ICA) are two methods for determining ER and PR. These two methods have not been compared with each other in relation to clinical endpoints. In the present study we prospectively evaluated the prognostic value of ER and PR as determined by ICA and EIA in 223 and 207 patients, respectively, with early breast cancer. ER was positive in approximately 77% of patients, while PR was positive in approximately 65% of patients. The proportion of potential agreement beyond chance between EIA and ICA was 0.58 and 0.65 for ER and PR, respectively. The median follow-up was 86 months. Both ER and PR appeared to be weak prognostic factors. There were no differences in prognostic value according to the time point of analysis or cutoff value chosen, nor were there any differences in the prognostic value of hormone receptors detected by ICA or EIA. Both methods appear to be equivalent in terms of qualification and prognostic value.
What is the finit state grammar of I C A ?
Rodger and Rosebrugh (1979) developed a method for identifying sets of rules by which discrete responses are combined to form functional sequences of behavior. In the present study, this so-called finite-state behavioral grammar method was used to analyze the food-hoarding behavior of Syrian golden hamsters (Mesocricetus auratus). An integral part of grammar derivation is the development of hypotheses about the function of the observed response patterns to guide the grammar derivation. Two such hypotheses guided the present analysis: first, we hypothesized that behaviors observed in the foraging area after the hamsters' cheek pouches were filled served an exploratory function; second, we hypothesized that behaviors observed in the home compartment prior to the hamsters emptying their cheeck pouches served to assess potential food deposition sites