Soft Clustering: Why and How-To

Rovetta, Stefano; Masulli, Francesco

doi:10.1007/978-3-030-12544-8_6

Cited by 2 publications

(6 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this work, we are particularly interested in applying some of our recent results in fuzzy clustering [42] for emotion recognition. The soft/fuzzy clustering approach consists in using a real-valued membership function instead of a categorical or binary membership decision.…”

Section: Clusteringmentioning

confidence: 99%

See 1 more Smart Citation

Emotion Recognition from Speech: An Unsupervised Learning Approach

Rovetta¹,

Mnasri²,

Masulli³

et al. 2020

IJCIS

Self Cite

View full text Add to dashboard Cite

Speech processing is quickly shifting toward affective computing, that requires handling emotions and modeling expressive speech synthesis and recognition. The latter task has been so far achieved by supervised classifiers. This implies a prior labeling and data preprocessing, with a cost that increases with the size of the database, in addition to the risk of committing errors. A typical emotion recognition corpus therefore has a relatively limited number of instances. To avoid the cost of labeling, and at the same time to reduce the risk of overfitting due to lack of data, unsupervised learning seems a suitable alternative to recognize emotions from speech. The recent advances in clustering techniques make it possible to reach good performances, comparable to that obtained by classifiers, with much less preprocessing load and even with generalization guarantees. This paper presents a novel approach for emotion recognition from speech signal, based on some variants of fuzzy clustering, such as probabilistic, possibilistic and graded-possibilistic fuzzy c-means. Experiments indicate that this approach (a) is effective in recognition, with in-corpus performances comparable to other proposals in the literature but with the added value of complexity control and (b) allows an innovative way to analyze emotions conveyed by speech using possibilistic membership degrees.

show abstract

Section: Clusteringmentioning

confidence: 99%

“…Also, it is worth noting that in case of probabilistic central clustering, ∑ c j=1 𝜇 jl = 1, whereas this condition is not necessarily met in the possibilistic and graded-possibilistic paradigms [42].…”

Section: The Graded Possibilistic C Means Algorithmmentioning

confidence: 99%

Emotion Recognition from Speech: An Unsupervised Learning Approach

Rovetta¹,

Mnasri²,

Masulli³

et al. 2020

IJCIS

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the particular case where v jl = e −d jl βj and β j > 0 (β j is a cluster width parameter selected a priory), the generalized partition function Z l is defined as In [15], the following parametrizations are suggested: α = 0 for fully possibilistic c-means α = 1 for fully probabilistic c-means 0 < α < 1 for graded-possibilistic c-means Besides, the cluster width parameter β j is suggested to be calculated as follows:…”

Section: Fuzzy Clusteringmentioning

confidence: 99%

“…Parameter setting Fuzzy clustering parameters were set based on results reported in [15] where it was proven that t = 1/2 guarantees no overlap between clusters, and where it was also observed that α should be close to 1. Then it was suggested to set α = log 2 (a + 1) 2 where a ∈ [0.5, 1] to have α ∈ [0.9, 1].…”

Section: Preprocessingmentioning

confidence: 99%

“…In this case, the function f defines the type of c-means, i.e.Z l = c j=1 v jl in case of probabilistic c-means Z l = 1 in case of possibilistic c-means Z l = ( c j=1 v jl ) α in case of graded-possibilistic c-meanswhere α ∈ [0, 1] is the degree of probabilistic tendency. Also, it should be noted that in case of probabilistic c-means, c j=1 µ jl = 1, whereas this condition is not necessarily met in possibilistic and graded-possibilistic c-means[15].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Emotion recognition from speech signal using fuzzy clustering

Rovetta¹,

Mnasri²,

Masulli³

et al. 2019

Proceedings of the 2019 Conference of the International Fuzzy Systems Association and the European Society for Fuzzy Logic and

Self Cite

View full text Add to dashboard Cite

Expressive speech modeling is a new trend in speech processing, including emotional speech synthesis and recognition. So far, emotion recognition from speech signal has been mainly achieved using supervised classifiers. However, clustering techniques seem well fitted to resolve such a problem, especially in huge databases, where speech labeling may be a hard and tedious task. This paper presents a novel approach for emotion recognition from speech signal, based on fuzzy clustering, including probabilistic, possibilistic and graded-possibilistic c-means. In comparison to crisp clustering, mainly using kmeans, fuzzy c-means look more fitted for this problem, and potentially offer an innovative way to analyze emotions conveyed by speech using membership degrees.

show abstract

Soft Clustering: Why and How-To

Cited by 2 publications

References 24 publications

Emotion Recognition from Speech: An Unsupervised Learning Approach

Emotion Recognition from Speech: An Unsupervised Learning Approach

Emotion recognition from speech signal using fuzzy clustering

Contact Info

Product

Resources

About