Classification of Imbalanced Data by Using the SMOTE Algorithm and Locally Linear Embedding

The future of Alberta's bitumen sands, also known as "oil sands" or "tar sands," and their place in Canada's energy future has become a topic of much public debate. Within this debate, the print, television, and social media campaigns of those who both support and oppose developing the oil sands are particularly visible. As such, campaigns around the oil sands may be seen as influencing audience perceptions of the benefits and drawbacks of oil sands production. There is consequently a need to study the media materials of various tar sands stakeholders and explore how they differ. In this setting, it is essential to gather documents and identify content within images, which requires the use of an image retrieval technique such as a content-based image retrieval (CBIR) system. In a CBIR system, images are represented by low-level features (i.e. specific structures in the image such as points, edges, or objects), which are used to distinguish pictures from one another.The oil sands domain has to date not been mapped using CBIR systems. The research thus focuses on creating an image retrieval system, namely Mediatoil-IR, for exploring documents related to the oil sands. Our aim is to evaluate various low-level representations of the images within this context. To this end, our experimental framework employs LAB color histogram (LAB) and speeded up robust features (SURF) in order to typify the imagery. We further use machine learning techniques to improve the quality of retrieval (in terms of both accuracy and speed). To achieve this aim, the extracted features from each image are encoded in the form of vectors and used as a training set for learning classification models to organize pictures into different categories. Different algorithms were considered such as Linear SVM, Quadratic SVM, Weighted KNN, Decision Trees, Bagging, and Boosting on trees. It was shown that Quadratic SVM algorithm trained on SURF features is a good approach for building CBIR, and is used in building Mediatoil-IR.Finally, with the help of created CBIR, we were able to extract the similar documents and explore the different types of imagery used by different stakeholders. Our experimental evaluation shows that our Mediatoil-IR system is able to accurately explore the imagery used by different stakeholders.iii

show abstract

“…This observation needs to be explored in our future work; we also plan to consider sampling methods [5,6,11,15,33,42,102].…”

Section: Figure 68 a Concept Showing Point Correspondence (A) Querymentioning

confidence: 99%

“…Under-sampling removes some instances of the majority class and thus may lead to a loss of information, whereas over-sampling generates artificial samples for the minority class. The various techniques for handling an imbalance are addressed in [5,6,11,15,33,42,102].…”

Section: Future Workmentioning

confidence: 99%

Exploring Mediatoil Imagery: A Content-based Approach

Saroop¹,

Viktor

McCurdy

et al. 2017

Proceedings of the 9th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management

View full text Add to dashboard Cite

show abstract

“…Among others, (Wang, 2008;Gao et al, 2011) can be listed. Moreover, several improvements of the original algorithm have been introduced (Chawla et al, 2003;Wang et al, 2006).…”

Section: Random Forests/smotementioning

confidence: 99%

Proceedings of the Third Workshop on Vision and Language

2014

View full text Add to dashboard Cite

PrefaceThe Workshop on Vision and Language 2014 (VL'14) took place in Dublin on 23rd July 2014, as part of COLING'14. It was the joint 3rd meeting of the EPSRC Network On Vision and Language and 1st technical meeting of the new European Network on Integrating Vision and Language which is funded as a European COST Action. The VL workshops have the general aims:1. to provide a forum for reporting and discussing planned, ongoing and completed research that involves both language and vision; and 2. to enable NLP and computer vision researchers to meet, exchange ideas, expertise and technology, and form new research partnerships.As funding for the V&L EPSRC Network (EP/H018557) ends and funding for the iV&L Net European COST Action (IC1307) starts, the focus of the VL workshops will shift onto integration and joint modelling of language and vision. iV&L Net will take over the organisation of annual VL workshops for the next four years as the flagship workshop of this new COST Action.The call for papers for VL'14 was issued in May 2014 and elicited a good number of highquality submissions, each of which was peer-reviewed by three members of the programme committee. The interest in the workshop from leading NLP and computer vision researchers and the quality of submissions was high, so we aimed to be as inclusive as possible within the practical constraints of the workshop. In the end we accepted 14 submissions as long papers, and eight as short papers.The resulting workshop programme packed a lot of exciting content into one day. We were delighted to be able to include in the programme a keynote presentation by Alex Jaimes of Yahoo! Inc., an internationally leading vision researcher. Our technical programme combined seven oral papers, seven long poster papers and seven short poster papers. Some thematic clusters emerged: combined text and image processing (Nguyen et al., Sakaki et al., Jones et al., Zhang et al., HaCohen-Kerner et al.), image description, annotation and labelling (Elliott, Liparas et al., Wang et al., Jokinen and Wilcock), data set creation (Weiland et al., Le et al., McGuinness et al.), situated dialogue (Summers-Stay et al., Schütte et al.), video analysis (Bhat and Olszewska, Shrestha et al.), aids for visually impaired people (Safi et al., Belz and Bharath), and visual analysis supported by text/speech features (Anbarjafari and Aabloo). The programme also included a discussion session on future directions for the VL community and workshops, including plans for shared task competitions.We would like to thank all the people who have contributed to the organisation and delivery of this workshop: the authors who submitted such high quality papers; the programme committee for their prompt and effective reviewing; our keynote speaker, Alex Jaimes; the COLING 2014 organising committee, especially the workshops chairs, Jennifer Foster, Dan Gildea, and Tim Baldwin; the participants in the workshop; and future readers of these proceedings for your shared interest in this exciting new area of research. Aug...

show abstract

“…The experimental results indicated that feature selections greatly enhanced the accuracy of classification and a particular feature selection using CART enhanced the classification accuracy of a particular datasets. Wang et al [6] proposed an over-sampling technique, LLE (the locally linear embedding algorithm) -based SMOTE (Synthetic Minority Oversampling TEechnique) to classify imbalanced dataset (chest x-ray image dataset). In the beginning, LLE algorithm is applied to handle the high-dimensional data into a low-dimensional space until the input variables are more separable, they oversampled data by SMOTE.…”

Section: Related Workmentioning

confidence: 99%

Duo Bundling Algorithms for Data Preprocessing: Case Study of Breast Cancer Data Prediction

Jojan¹,

Srivihok²

2014

LNSE

View full text Add to dashboard Cite

Abstract-Classification of imbalanced dataset is the most popular and challenged problems for researchers to solve in nowadays. This paper proposed a two-steps approach to improve the quality of class prediction imbalanced breast cancer dataset. The two-steps approach consists of two main techniques: 1) using feature selection techniques to filter out unimportant features from the dataset; and 2) using the over-sampling technique to adjust the size of the minority class to be similar to the size of the majority class. The three different classification algorithms: artificial neural network (MLP), decision tree (C4.5) and Naï ve Bayes, were applied. The classification result indicated that C4.5 was the most suitable to classify this dataset which can give the highest accuracy of 83.80%.

show abstract

Classification of Imbalanced Data by Using the SMOTE Algorithm and Locally Linear Embedding

Cited by 148 publications

References 3 publications

Exploring Mediatoil Imagery: A Content-based Approach

Exploring Mediatoil Imagery: A Content-based Approach

Proceedings of the Third Workshop on Vision and Language

Duo Bundling Algorithms for Data Preprocessing: Case Study of Breast Cancer Data Prediction

Contact Info

Product

Resources

About