DOCSLIB.ORG

Text Similarity Estimation Based on Word Embeddings and Matrix Norms for Targeted Marketing

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Text Similarity Estimation Based on Word Embeddings and Matrix Norms for Targeted Marketing Text Similarity Estimation Based on Word Embeddings and Matrix Norms for Targeted Marketing Tim vor der Bruck¨ Marc Pouly School of Information Technology School of Information Technology Lucerne University of Lucerne University of Applied Sciences and Arts Applied Sciences and Arts Switzerland Switzerland [email protected] [email protected] Abstract determined beforehand on a very large cor- pus typically using either the skip gram or The prevalent way to estimate the similarity of two documents based on word embeddings the continuous bag of words variant of the is to apply the cosine similarity measure to Word2Vec model (Mikolov et al., 2013). the two centroids obtained from the embed- 2. The centroid over all word embeddings be- ding vectors associated with the words in each longing to the same document is calculated document. Motivated by an industrial appli- to obtain its vector representation. cation from the domain of youth marketing, If vector representations of the two documents to where this approach produced only mediocre compare were successfully established, a similar- results, we propose an alternative way of com- ity estimate can be obtained by applying the cosine bining the word vectors using matrix norms. The evaluation shows superior results for most measure to the two vectors. of the investigated matrix norms in compari- Let x1; : : : ; xm and y1; : : : ; yn be the word vec- son to both the classical cosine measure and tors of two documents. The cosine similarity value several other document similarity estimates. between the two document centroids C1 und C2 is given by: 1 Introduction Estimating semantic document similarity is of ut- m n 1 X 1 X most importance in a lot of different areas, like cos( ( x ; y )) \ m i n i plagiarism detection, information retrieval, or text i=1 i=1 (1) Pm Pn summarization. We will focus here on an NLP hxi; yji = i=1 j=1 application that has been less researched, i.e., the mnkC1kkC2k assignment of people to the best matching target group to allow for running precise and customer- Hence, potentially small values of hxi; yji can oriented marketing campaigns. have in aggregate a considerable influence on the Until recently, similarity estimates were pre- total similarity estimate, which makes this esti- dominantly based either on deep semantic ap- mate vulnerable to noise in the data. We propose proaches or on typical information retrieval tech- an alternative approach that is based on matrix niques like Latent Semantic Analysis. In the last norms and which proved to be more noise-robust couple of years, however, so-called word and sen- by focusing primarily on high word similarities. tence embeddings became state-of-the-art. Finally, we conducted an evaluation where we The prevalent approach to document similarity achieved with our method superior accuracy in estimation based on word embeddings consists in target group assignments than several traditional measuring the similarity between the vector rep- word embedding based methods. resentations of the two documents derived as fol- 2 Related Work lows: 1. The word embeddings (often weighted by The most popular method to come up with word the tf-idf coefficients of the associated words vectors is Word2Vec, which is based on a 3 layer (Brokos et al., 2016)) are looked up in a neural network architecture in which the word hashtable for all the words in the two doc- vectors are obtained as the weights of the hid- uments to compare. These embeddings are den layer. Alternatives to Word2Vec are GloVe 1827 Proceedings of NAACL-HLT 2019, pages 1827–1836 Minneapolis, Minnesota, June 2 - June 7, 2019. c 2019 Association for Computational Linguistics (Pennington et al., 2014), which is based on ag- Name Definition gregated global word co-occurrence statistics and qPm Pn 2 the Explicit Semantic Analysis (or shortly ESA) Frob. norm kAkF := i=1 j=1 jAijj p > (Gabrilovic and Markovitch, 2009), in which each 2-norm kAk2 := ρ(A A) word is represented by the column vector in the Pm Pn L1;1-norm kAkL1;1 := i=1 j=1 jAijj tf-idf matrix over Wikipedia. Pm 1-norm kAk1 := max1≤j≤n i=1 jAijj Pn The idea of Word2Vec can be transferred to the 1-norm kAk1 := max1≤i≤m j=1 jAijj level of sentences as well. In particular, the so- called Skip Thought Vector model (STV) (Kiros Table 1: Examples of matrix norms; A is an m × n et al., 2015) derives a vector representation of matrix; ρ(X) denotes the largest absolute eigenvalue of a squared matrix X. the current sentence by predicting the surrounding sentences. (Song and Roth, 2015) propose an alternative approach to applying the cosine measure to the ponential runtime complexity. two word vector centroids for ESA word embed- A drawback of conventional similarity esti- dings. In particular, they establish a bipartite graph mates as described above is that slightly related consisting of the best matching vector compo- word pairs can have in aggregate a considerable nents by solving a linear optimization problem. influence on their values, i.e., these estimates are The similarity estimate for the documents is then sensitive to noise in the data. In contrast, several given by the global optimum of the objective func- of our matrix norm based similarity estimates fo- tion. However, this method is only useful for cus primarily on strongly related word pairs and sparse vector representations. In case of dense are therefore less vulnerable to noise. vectors, (Mijangos et al., 2017) suggested to ap- ply the Frobenius kernel to the embedding matri- ces, which contain the embedding vectors for all 3 Similarity Measure / Matrix Norm document components (usually either sentences or words) (cf. also (Hong et al., 2015)). However, crucial limitations are that the Frobenius kernel Before going more into detail, we want to review is only applicable if the number of words (sen- some concepts that are crucial for the remainder tences respectively) in the compared documents of this paper. According to (Belanche and Orozco, coincide and that a word from the first document is 2011), a similarity measure on some set X is only compared with its counterpart from the sec- an upper bounded, exhaustive and total function ond document. Thus, an optimal matching has to s : X × X ! I ⊂ R with jIj > 1 (therefore be established already beforehand. In contrast, the I is upper bounded and sup I exists). Addition- matrix norm approach as presented here applies to ally, it should fulfill the properties of reflexivity arbitrary embedding matrices. Since it conducts (the supremum is reached if an item is compared a pairwise comparison of all words contained in to itself) and symmetry. We call such a measure the two documents, there is also no need for any normalized if the supremum equals 1 (Attig and matching method. Perner, 2011). Note that an asymmetric similarity measure can easily be converted into a symmet- Another simlarity estimate that employs the en- ric by taking the geometric or arithmetic mean of tire embedding matrix is the word mover‘s dis- the asymmetric measure applied twice to the same tance (Kusner et al., 2015), which is a special case arguments in switched order. of the earth mover‘s distance, a well studied trans- portation problem. Basically, this approach deter- A norm is a function f : V ! R over some vec- mines the minimum effort (with respect to em- tor space V that is absolutely homogeneous, pos- bedding vector changes) to transform the words itive definite and fulfills the triangle inequality. It of one text into the words of another text. The is called matrix norm if its domain is a set of ma- word mover‘s distance requires a linear optimiza- trices and if it is sub-multiplicative, i.e., kABk ≤ tion problem to be solved. Linear optimization is kAk·kBk. Several popular matrix norms are given usually tackled by the simplex method, which has in Table1. Note that the Frobenius norm can also p > in the worst case, which rarely occurs however, ex- be represented by kAkF = tr(AA ). 1828 4 Document Similarity Measure based for arbitrary matrices Z, since with this property on Matrix Norms we have kA>Bk For an arbitrary document t we define the em- sn(t; u) = pkA>Ak · kB>Bk beddings matrix E(t) as follows: E(t)ij is the i- th component of the normalized embeddings vec- k(B>A)>k = p (2) tor belonging to the j-th word of the document t. kB>Bk · kA>Ak Let t; u be two arbitrary documents, then the en- kB>Ak = p = sn(u; t) try (i; j) of a product E(t)>E(u) specifies the re- kB>Bk · kA>Ak sult of the cosine measure estimating the seman- Let M and N be arbitrary matrices such that tic similarity between word i of document t and MN and NM are both defined and quadratic, word j of document u. The value of a matrix norm then (see (Chatelin, 1993)) kE(t)>E(u)k is then a measure for the similar- ity of the two documents. Since the vector com- ρ(MN) = ρ(NM) (3) ponents obtained by Word2Vec can be negative, where ρ(X) denotes the largest absolute eigen- the cosine measure between two word vectors can value of a squared matrix X. also assume negative values (rather rarely in prac- Using identity3 one can easily infer that: tice though). Negative cosine values indicate neg- q q > > > atively correlated words and should be handled kZk2 = ρ(Z Z) = ρ(ZZ ) = kZ k2 akin to the uncorrelated case. Because a matrix (4) norm usually treats negative and positive matrix entries alike, we replace all negative values in the matrix by zeros.
Load more

Recommended publications
  • Soft Similarity and Soft Cosine Measure: Similarity of Features in Vector Space Model
    Soft Similarity and Soft Cosine Measure: Similarity of Features in Vector Space Model Grigori Sidorov1, Alexander Gelbukh1, Helena Gomez-Adorno1, and David Pinto2 1 Centro de Investigacion en Computacion, Instituto Politecnico Nacional, Mexico D.F., Mexico 2 Facultad de Ciencias de la Computacion, Benemerita Universidad Autonoma de Puebla, Puebla, Mexico {sidorov,gelbukh}@cic.ipn.mx, [email protected], [email protected] Abstract. We show how to consider similarity between 1 Introduction features for calculation of similarity of objects in the Vec­ tor Space Model (VSM) for machine learning algorithms Computation of similarity of specific objects is a basic and other classes of methods that involve similarity be­ task of many methods applied in various problems in tween objects. Unlike LSA, we assume that similarity natural language processing and many other fields. In between features is known (say, from a synonym dictio­ natural language processing, text similarity plays crucial nary) and does not need to be learned from the data. role in many tasks from plagiarism detection [18] and We call the proposed similarity measure soft similarity. question answering [3] to sentiment analysis [14-16]. Similarity between features is common, for example, in The most common manner to represent objects is natural language processing: words, n-grams, or syn­ the Vector Space Model (VSM) [17]. In this model, the tactic n-grams can be somewhat different (which makes objects are represented as vectors of values of features. them different features) but still have much in common: The features characterize each object and have numeric for example, words “play” and “game” are different but values.
    [Show full text]
  • Evaluating Vector-Space Models of Word Representation, Or, the Unreasonable Effectiveness of Counting Words Near Other Words
    Evaluating Vector-Space Models of Word Representation, or, The Unreasonable Effectiveness of Counting Words Near Other Words Aida Nematzadeh, Stephan C. Meylan, and Thomas L. Griffiths University of California, Berkeley fnematzadeh, smeylan, tom griffi[email protected] Abstract angle between word vectors (e.g., Mikolov et al., 2013b; Pen- nington et al., 2014). Vector-space models of semantics represent words as continuously-valued vectors and measure similarity based on In this paper, we examine whether these constraints im- the distance or angle between those vectors. Such representa- ply that Word2Vec and GloVe representations suffer from the tions have become increasingly popular due to the recent de- same difficulty as previous vector-space models in capturing velopment of methods that allow them to be efficiently esti- mated from very large amounts of data. However, the idea human similarity judgments. To this end, we evaluate these of relating similarity to distance in a spatial representation representations on a set of tasks adopted from Griffiths et al. has been criticized by cognitive scientists, as human similar- (2007) in which the authors showed that the representations ity judgments have many properties that are inconsistent with the geometric constraints that a distance metric must obey. We learned by another well-known vector-space model, Latent show that two popular vector-space models, Word2Vec and Semantic Analysis (Landauer and Dumais, 1997), were in- GloVe, are unable to capture certain critical aspects of human consistent with patterns of semantic similarity demonstrated word association data as a consequence of these constraints. However, a probabilistic topic model estimated from a rela- in human word association data.
    [Show full text]
  • Using Semantic Folding with Textrank for Automatic Summarization
    DEGREE PROJECT IN COMPUTER SCIENCE AND ENGINEERING, SECOND CYCLE, 30 CREDITS STOCKHOLM, SWEDEN 2017 Using semantic folding with TextRank for automatic summarization SIMON KARLSSON KTH ROYAL INSTITUTE OF TECHNOLOGY SCHOOL OF COMPUTER SCIENCE AND COMMUNICATION Using semantic folding with TextRank for automatic summarization SIMON KARLSSON [email protected] Master in Computer Science Date: June 27, 2017 Principal: Findwise AB Supervisor at Findwise: Henrik Laurentz Supervisor at KTH: Stefan Nilsson Examiner at KTH: Olov Engwall Swedish title: TextRank med semantisk vikning för automatisk sammanfattning School of Computer Science and Communication i Abstract This master thesis deals with automatic summarization of text and how semantic fold- ing can be used as a similarity measure between sentences in the TextRank algorithm. The method was implemented and compared with two common similarity measures. These two similarity measures were cosine similarity of tf-idf vectors and the number of overlapping terms in two sentences. The three methods were implemented and the linguistic features used in the construc- tion were stop words, part-of-speech filtering and stemming. Five different part-of- speech filters were used, with different mixtures of nouns, verbs, and adjectives. The three methods were evaluated by summarizing documents from the Document Understanding Conference and comparing them to gold-standard summarization cre- ated by human judges. Comparison between the system summaries and gold-standard summaries was made with the ROUGE-1 measure. The algorithm with semantic fold- ing performed worst of the three methods, but only 0.0096 worse in F-score than cosine similarity of tf-idf vectors that performed best. For semantic folding, the average preci- sion was 46.2% and recall 45.7% for the best-performing part-of-speech filter.
    [Show full text]
  • Lecture Note
    Algorithms for Data Science: Lecture on Finding Similar Items Barna Saha 1 Finding Similar Items Finding similar items is a fundamental data mining task. We may want to find whether two documents are similar to detect plagiarism, mirror websites, multiple versions of the same article etc. Finding similar items is useful for building recommender systems as well where we want to find users with similar buying patterns. In Netflix two movies can be deemed similar if they are rated highly by the same customers. While, there are many measures of similarity, in this lecture, we will concentrate one such popular measure known as Jaccard Similarity. Definition (Jaccard Similairty). Given two sets S1 and S2, Jaccard similarity of S1 and S2 is defined as |S1∩S2 S1∪S2 Example 1. Let S1 = {1, 2, 3, 4, 7} and S2 = {1, 4, 9, 7, 5} then |S1 ∪ S2| = 7 and |S1 ∩ S2| = 3. 3 Thus the Jaccard similarity of S1 and S2 is 7 . 1.1 Document Similarity To compare two documents to know how similar they are, here is a simple approach: • Compute k shingles for a suitable value of k k shingles are all substrings of length k that appear in the document. For example, if a document is abcdabd and k = 2, then the 2-shingles are ab, bc, cd, da, ab, bd. • Compare the set of k shingled based on Jaccard similarity. One will often map the set of shingled to a set of integers via hashing. What should be the right size of the hash table? Select a hash table size to avoid collision.
    [Show full text]
  • Learning Semantic Similarity for Very Short Texts
    Learning Semantic Similarity for Very Short Texts Cedric De Boom, Steven Van Canneyt, Steven Bohez, Thomas Demeester, Bart Dhoedt Ghent University – iMinds Gaston Crommenlaan 8-201, 9050 Ghent, Belgium fcedric.deboom, steven.vancanneyt, steven.bohez, thomas.demeester, [email protected] Abstract—Levering data on social media, such as Twitter a single sentence representation that contains most of its and Facebook, requires information retrieval algorithms to semantic information. Many authors choose to average or become able to relate very short text fragments to each maximize across the embeddings in a text [8], [9], [10] or other. Traditional text similarity methods such as tf-idf cosine- similarity, based on word overlap, mostly fail to produce combine them through a multi-layer perceptron [6], [11], by good results in this case, since word overlap is little or non- clustering [12], or by trimming the text to a fixed length existent. Recently, distributed word representations, or word [11]. embeddings, have been shown to successfully allow words The Paragraph Vector algorithm by Le and Mikolov— to match on the semantic level. In order to pair short text also termed paragraph2vec—is a powerful method to find fragments—as a concatenation of separate words—an adequate distributed sentence representation is needed, in existing lit- suitable vector representations for sentences, paragraphs and erature often obtained by naively combining the individual documents of variable length [13]. The algorithm tries to word representations. We therefore investigated several text find embeddings for separate words and paragraphs at the representations as a combination of word embeddings in the same time through a procedure similar to word2vec.
    [Show full text]
  • Word Embeddings for Similarity Scoring in Practical Information
    Names of Editors (Hrsg.): 47. Jahrestagung der Gesellschaft für Informatik e.V. (GI), Lecture Notes in Informatics (LNI), Gesellschaft für Informatik, Bonn 2017 1 Evaluating the Impact of Word Embeddings on Similarity Scoring in Practical Information Retrieval Lukas Galke12, Ahmed Saleh12, Ansgar Scherp12 Abstract: We assess the suitability of word embeddings for practical information retrieval scenarios. Thus, we assume that users issue ad-hoc short queries where we return the first twenty retrieved documents after applying a boolean matching operation between the query and the documents. We compare the performance of several techniques that leverage word embeddings in the retrieval models to compute the similarity between the query and the documents, namely word centroid similarity, paragraph vectors, Word Mover’s distance, as well as our novel inverse document frequency (IDF) re-weighted word centroid similarity. We evaluate the performance using the ranking metrics mean average precision, mean reciprocal rank, and normalized discounted cumulative gain. Additionally, we inspect the retrieval models’ sensitivity to document length by using either only the title or the full-text of the documents for the retrieval task. We conclude that word centroid similarity is the best competitor to state-of-the-art retrieval models. It can be further improved by re-weighting the word frequencies with IDF before aggregating the respective word vectors of the embedding. The proposed cosine similarity of IDF re-weighted word vectors is competitive to the TF-IDF baseline and even outperforms it in case of the news domain with a relative percentage of 15%. Keywords: Word embeddings; Document representation; Information retrieval 1 Introduction Word embeddings have become the default representation for text in many neural network architectures and text processing pipelines [BCV13; Be03; Go16].
    [Show full text]
  • Correlation Coefficients and Semantic Textual Similarity
    Correlation Coefficients and Semantic Textual Similarity Vitalii Zhelezniak, Aleksandar Savkov, April Shen & Nils Y. Hammerla Babylon Health ffirstname.lastnameg @babylonhealth.com Abstract text, with methods ranging from clever composi- tions of word embeddings (Mitchell and Lapata, A large body of research into semantic tex- 2008; De Boom et al., 2016; Arora et al., 2017; tual similarity has focused on constructing Wieting et al., 2016; Wieting and Gimpel, 2018; state-of-the-art embeddings using sophisti- Zhelezniak et al., 2019) to sophisticated neural ar- cated modelling, careful choice of learning signals and many clever tricks. By contrast, chitectures (Le and Mikolov, 2014; Kiros et al., little attention has been devoted to similar- 2015; Hill et al., 2016; Conneau et al., 2017; Gan ity measures between these embeddings, with et al., 2017; Tang et al., 2017; Zhelezniak et al., cosine similarity being used unquestionably 2018; Subramanian et al., 2018; Pagliardini et al., in the majority of cases. In this work, we 2018; Cer et al., 2018). illustrate that for all common word vectors, Comparatively, there is little research into sim- cosine similarity is essentially equivalent to the Pearson correlation coefficient, which pro- ilarity measures for textual embeddings. Despite vides some justification for its use. We thor- some investigations into alternatives (Camacho- oughly characterise cases where Pearson cor- Collados et al., 2015; De Boom et al., 2015; Santus relation (and thus cosine similarity) is unfit as et al., 2018; Zhelezniak et al., 2019), cosine simi- similarity measure. Importantly, we show that larity has persistently remained the default and un- Pearson correlation is appropriate for some questioned choice across the field.
    [Show full text]
  • A Geometric Interpretation for Local Alignment-Free Sequence Comparison
    JOURNAL OF COMPUTATIONAL BIOLOGY Volume 20, Number 7, 2013 Research Articles # Mary Ann Liebert, Inc. Pp. 471–485 DOI: 10.1089/cmb.2012.0280 A Geometric Interpretation for Local Alignment-Free Sequence Comparison EHSAN BEHNAM, MICHAEL S. WATERMAN, and ANDREW D. SMITH ABSTRACT Local alignment-free sequence comparison arises in the context of identifying similar seg- ments of sequences that may not be alignable in the traditional sense. We propose a ran- domized approximation algorithm that is both accurate and efficient. We show that under à D2 and its important variant D2 as the similarity measure, local alignment-free comparison between a pair of sequences can be formulated as the problem of finding the maximum bichromatic dot product between two sets of points in high dimensions. We introduce a geometric framework that reduces this problem to that of finding the bichromatic closest pair (BCP), allowing the properties of the underlying metric to be leveraged. Local align- ment-free sequence comparison can be solved by making a quadratic number of alignment- free substring comparisons. We show both theoretically and through empirical results on simulated data that our approximation algorithm requires a subquadratic number of such comparisons and trades only a small amount of accuracy to achieve this efficiency. There- fore, our algorithm can extend the current usage of alignment-free–based methods and can also be regarded as a substitute for local alignment algorithms in many biological studies. Key words: algorithms, alignment, dynamic programming, metagenomics. 1. INTRODUCTION equence alignment is perhaps the most well-known and intensively studied computational problem in Smodern molecular biology.
    [Show full text]
  • Chapter 6: Vector Semantics What Do Words Mean?
    Dan Jurafsky and James Martin Speech and Language Processing Chapter 6: Vector Semantics What do words mean? First thought: look in a dictionary http://www.oed.com/ Oxford English Dictionary | The definitive record of the English language pepperWords,, n. Lemmas, Senses, Definitions Pronunciation: Brit. /ˈpɛpə/ , U.S. /ˈpɛpər/ OxfordForms: OE peoporEnglish (rare), OEDictionary pipcer (transmission | The error), OEdefinitive pipor, OE pipur record (rare ... of the English languageFrequency (in current use): Etymology: A borrowing from Latin. Etymon: Latin piper. < classical Latin piper, a loanword < Indo-Aryan (as is ancient Greek πέπερι ); compare Sanskritsense ... lemma betel-, malagueta, walldefinition pepper, etc.: see the first element. See also WATER PEPPER n. 1. pepper, n. betel-, malagueta, wall pepper, etc.: see the first element. See also WATER PEPPER n. 1. I. The spice or the plant. Pronunciation: 1. Brit. /ˈpɛpə/ , U.S. /ˈpɛpər/ Forms: OE peopor (rare), OE pipcer (transmission error), OE pipor, OE pipur (rare ... a. A hot pungent spice derived from the prepared fruits (peppercorns) of c. U.S. The California pepper tree, Schinus molle. Cf. PEPPER TREE n. 3. theFrequency pepper plant,(in current Piper use):nigrum (see sense 2a), used from early times to c. U.S. The California pepper tree, Schinus molle. Cf. PEPPER TREE n. 3. seasonEtymology: food, Aeither borrowing whole from or groundLatin. Etymon: to powder Latin (often piper .in association with salt).< classical Also Latin (locally, piper chiefly, a loanword with <distinguishing Indo-Aryan (as word):is ancient a similarGreek πέπερι spice ); compare Sanskrit ... derived from the fruits of certain other species of the genus Piper; the 3.
    [Show full text]
  • Breaking and Fixing Secure Similarity Approximations: Dealing with Adversarially Perturbed Inputs
    Breaking and Fixing Secure Similarity Approximations: Dealing with Adversarially Perturbed Inputs Evgenios M. Kornaropoulos Petros Efstathopoulos Dept. of Computer Science, Brown University Symantec Research Labs [email protected] petros [email protected] Abstract—Computing similarity between data is a fundamen- working with a succinct representation of the data that is tal problem in information retrieval and data mining. To called a sketch. address the relevant performance and scalability challenges, Sketching is the mainstream approach for efficiently approximation methods are employed for large-scale similarity approximating a plethora of functions [11], [12], [16], [18], computation. A common characteristic among all privacy- [25], [36], [42], [47], [57], [58]. The seminal work by preserving approximation protocols based on sketching is that Feigenbaum et al. [27] set the foundation for secure multi- the sketching is performed locally and is based on common party computation of approximation functions. Since then, randomness. researchers proposed a variety of secure approximations In the semi-honest model the input to the sketching algo- based on sketching for several application areas. Melis et rithm is independent of the common randomness. We, however, al. [45] propose scalable and secure protocols to approx- consider a new threat model where a party is allowed to imate statistics on large data streams. Blundo et al. [9] use the common randomness to perturb her input 1) offline, propose secure approximation algorithms for iris matching, and 2) before the execution of any secure protocol so as to multimedia content similarity, and document similarity. The steer the approximation result to a maliciously chosen output. community has made several important steps towards private We formally define perturbation attacks under this adversarial computation on genomic data in a time-efficient and scalable model and propose two attacks on the well-studied techniques manner [3], [17], [22], [50].
    [Show full text]
  • Experimental Analysis of Crisp Similarity and Distance Measures
    International Conference of Soft Computing and Pattern Recognition Experimental Analysis of Crisp Similarity and Distance Measures Leila Baccour Robert I. John REGIM-Lab.: REsearch Groups in Intelligent Machines, Automated Scheduling, University of Sfax, Optimisation and Planning (ASAP) group, ENIS, BP 1173, Sfax, 3038, Tunisia. Computer Science Email: [email protected] Jubilee Campus, Wollaton Road Nottingham, NG8 1BB, UK Email: [email protected] Abstract—Distance measures are a requirement in many clas- This function is called distance, index of distance, index of sification systems. Euclidean distance is the most commonly used dissimilarity, semi-metric distance, or distance ultra-metric in these systems. However, there exists a huge number of distance according to one of the following cases: and similarity measures. In this paper we present an experimental index of dissimilarity: if the function d satisfies the analysis of crisp distance and similarity measures applied to • shapes classification and Arabic sentences recognition. properties 1 and 2. index of distance or semi-metric distance: if the function Keywords-Similarity measures, distance measures, crisp sets, • classification of shapes, classification of Arabic sentences d satisfies the properties 1, 2 and 3 distance: if the function d satisfies the properties 1, 2, 3 • I. INTRODUCTION and 4 distance ultra-metric: if the function d verifies the prop- In some applications such as recognition, classification and • clustering, it is necessary to compare two objects described erty x, y, z E,d(x, y) max (d (x, z) ,d(z,y)) ∀ ∈ ≤ with vectors of features. This operation is based on a distance Two patterns x and y are close if the distance d(x, y) tends or a similarity measure.
    [Show full text]
  • Textual Similarity
    Bachelor Thesis in Textual Similarity Abdulla Ali Kongens Lyngby IMM-BSc-2011-19 P a g e | 2 Technical University of Denmark Informatics and Mathematical Modelling Building 321, DK-2800 Kongens Lyngby, Denmark Phone +45 45253351, Fax +45 45882673 [email protected] www.imm.dtu.dk IMM-BSc: ISSN 2011-19 P a g e | 3 Abstract This thesis deals with one of Information Retrieval’s big interest: Textual Similarity. Textual Similarity is a process where two texts are compared to find the Similarity between them. This thesis goes into depth with subjects as Similarity and its definition, Language and Words, then algorithms and other tools used to find Textual Similarity. Two algorithms, Optimal String Alignment Distance and Cosine Similarity, are then implemented with and without Stemmer and Stop Word removal algorithms. Many tests with the structures of the sentences being different from each other have been run. These have then been compared to the algorithm options to decide which one works best. These have also been compared with the writer’s expectations and human testers. The result was that while Optimal String Alignment Distance had its advantages in some fields, Cosine Similarity had its own in others. It was also concluded that both run better with a stemmer and stop word removal added to them than without. P a g e | 4 Abstrakt Denne opgave handler om en af Information Retrieval’s store interesse: Tekstsammenligning. Tekstsammenligning er en process hvor to tekster sammenlignings for at finde deres lighed. Denne opgave will gå i dybden med emner som lighed og dennes definition, sprog og ord,så algoritmer og andre værktøjer for at finde tekstsammenligning.
    [Show full text]