| 1. |
- Gyllenberg, Mats, et al.
(författare)
-
Non-uniqueness in probabilistic numerical identification of bacteria
- 1994
-
Ingår i: Journal of Applied Probability. - : Cambridge University Press (CUP). - 0021-9002 .- 1475-6072. ; 31:2, s. 542-548
-
Tidskriftsartikel (refereegranskat)abstract
- In this note we point out an inherent difficulty in numerical identification of bacteria. The problem is that of uniqueness of the taxonomic structure or, in mathematical terms, the lack of statistical identifiability of finite mixtures of multivariate Bernoulli probability distributions shown here.
|
|
| 2. |
- Gyllenberg, Mats, et al.
(författare)
-
Non-uniqueness of numerical taxonomic structures
- 1993
-
Ingår i: Binary Computing in Microbiology. - 0266-304X. ; 5:4, s. 138-144
-
Tidskriftsartikel (refereegranskat)abstract
- The most important methods of numerical taxonomy in microbiology are based on so called reference matrices giving the frequencies of positive binary features of the different taxa. Microbiologists seem to have been tacitly assuming that every well-defined classification method, that is, every algorithm for constructing a reference matrix from data, leads to a unique classification (reference matrix). We use a mathematical result-that a finite mixture of multivariate Bernoulli distributions is always unidentifiable-to disprove this assumption. We show that the same classification method applied to the same data can always lead to different classifications. This result is of importance for the foundations of computational microbial taxonomy. It is illustrated by simple examples from the two main methods of classification and identification: the one where classification is performed first and then followed by identification, and cumulative classification where classification and identification are carried out simultaneously. The consequences of the non-uniqueness result for microbiological practice are discussed
|
|
| 3. |
- Gyllenberg, Mats, et al.
(författare)
-
Null recurrence in a stochastic Ricker model
- 1994
-
Ingår i: Analysis, algebra, and computers in mathematical research. - New York : Marcel Dekker Incorporated. - 0824792173 ; , s. 147-164
-
Konferensbidrag (refereegranskat)abstract
- We consider a nonlinear first order stochastic difference equation which may be viewed as a stochastic perturbation of {\it W. E. Ricker's} [J. Fish. Res. Bd. Can. 11, 559-623 (1954)] deterministic model of population growth. Numerical experiments seem to suggest that the corresponding Markov process has a stationary probability distribution but this is shown to be false by proving that the process is in fact null recurrent
|
|
| 4. |
- Gyllenberg, Mats, et al.
(författare)
-
Population models with environmental stochasticity
- 1994
-
Ingår i: Journal of Mathematical Biology. - 0303-6812 .- 1432-1416. ; 32:2, s. 93-108
-
Tidskriftsartikel (refereegranskat)abstract
- Two discrete population models, one with stochasticity in the carrying capacity and one with stochasticity in the per capita growth rate, are investigated. Conditions under which the corresponding Markov processes are null recurrent and positively recurrent are derived
|
|
| 5. |
- Janzura, M., et al.
(författare)
-
Minimum entropy of error principle in estimation
- 1994
-
Ingår i: Information Sciences. - : Elsevier BV. - 0020-0255 .- 1872-6291. ; 79:1-2, s. 123-144
-
Tidskriftsartikel (refereegranskat)abstract
- The principle of minimum error entropy estimation as found in the work of Weidemann and Stear is reformulated as a problem of finding optimum locations of probability densities in a given mixture such that the resulting (differential) entropy is minimized. New results concerning the entropy lower bound are derived. Continuity of the entropy and attaining the minimum entropy are proved in the case where the mixture is finite. Some other examples and situations, in particular that of symmetric unimodal densities, are studied in more detail
|
|
| 6. |
|
|
| 7. |
- Koski, Timo
(författare)
-
A diffusion approximation for the quantization error in delta modulation with a Gauss-Markov signal
- 1990
-
Ingår i: Limit theorems in probability and statistics. - Amsterdam : Elsevier. - 0444987584 ; , s. 305-325
-
Bokkapitel (övrigt vetenskapligt/konstnärligt)abstract
- A stochastic differential equation (SDE) for approximation in the sense of weak convergence of the error process in delta modulation for a Gauss-Markov process is established. The pertinent SDE has been previously proposed by E. N. Protonotarios. Here we use some results about the representation of the encoded process in terms of a stochastic integral and the weak convergence methods of H. Kushner.
|
|
| 8. |
- Koski, Timo
(författare)
-
A note on a probabilistic decomposition of linear delta modulator of a Wiener process
- 1988
-
Ingår i: Annales Academiae Scientiarum Fennicae. Series A. I, Mathematica. - 0066-1953. ; 13:2, s. 209-218
-
Tidskriftsartikel (refereegranskat)abstract
- Linear delta modulation (LDM) of a standard Wiener process is examined and it is shown that the encoder can be interpreted in terms of a functional of the predicted reconstruction error and a stochastic integral. The martingale type structure associated with the decoder process is studied and some of the properties are extended to other diffusion sources. The Malliavin calculus is applied to study further the predicted error in the encoded process and it turns out that this part possesses a probability density function.
|
|
| 9. |
- Koski, Timo, et al.
(författare)
-
Clustering by Adaptive Local Search with multiple search operators
- 2000
-
Ingår i: Pattern Analysis and Applications. - : Springer Science and Business Media LLC. - 1433-7541 .- 1433-755X. ; 3:4, s. 348-357
-
Tidskriftsartikel (refereegranskat)abstract
- Local Search (LS) has proven to be an efficient optimisation technique in clustering applications and in the minimisation of stochastic complexity of a data set. In the present paper, we propose two ways of organising LS in these contexts, the Multi-operator Local Search (MOLS) and the Adaptive Multi-Operator Local Search (AMOLS), and compare their performance to single operator (random swap) LS method and repeated GLA (Generalised Lloyd Algorithm). Both of the proposed methods use several different LS operators to solve the problem. MOLS applies the operators cyclically in the same order, whereas AMOLS adapts itself to favour the operators which manage to improve the result more frequently. We use a large database of binary vectors representing strains of bacteria belonging to the family Enterobacteriaceae and a binary image as our test materials. The new techniques turn out to be very promising in these tests.
|
|
| 10. |
- Koski, Timo
(författare)
-
Nonlinear autoregression in the theory of signal compression
- 1992
-
Ingår i: Annales Academiae Scientiarum Fennicae. Series A. I, Mathematica. - 0066-1953. ; 17:1, s. 51-64
-
Tidskriftsartikel (refereegranskat)abstract
- The nonlinear autoregression Un+1=Un-b·sign(Un)+m+n, where ξn is a white noise, emerges both in delta modulation and in sigma-delta modulation. The author proves the existence of a unique invariant measure and some ergodic properties, and discusses a time continuous approximation of the process
|
|
