Research Article Article Page 174 Journal ofof ResearchResearch and and Review Review in inScience, Science, 174–183 174–183 Volume 4, December 2017 DOI:10.36108/jrrslasu/7102/40(0152)Volume 4, December 2017 ORIGINAL RESEARCH

PERFORMANCE EVALUATION OF ANAGRAM ORTHOGRPHIC SIMILARITY MEASURES

Raji-Lawal Hanat Y.1,2, Akinwale Adio T.2 ,Folorunsho O.2, and Mustapha Amidu O.2

1Deparment of Computer, cience, Abstract: Lagos State University, Ojo, Lagos Introduction: Anagram solving task involves a retrieving process from previously acquired knowledge, this serves as a suitable memory cognition 2Department of Computer, test. Automation of this process can give a very good memory cognitive Science, Federal University Of Agriculture Abeokuta test tool, the method behind this automation is anagram orthographic similarity measure. Correspondence Aim: The purpose of this research is to study existing anagram Raji-Lawal Hanat Y., Department of Computer Science, Faculty of orthographic similarity measures, to deduce their strengths and Science, Lagos State University weaknesses, for further improvement. Ojo, Lagos, Nigeria Materials and Methods: Experiments were carried out on the measures, Email:[email protected] using real data. Their behaviour on different orthographic string set was observed. Result: Experiment revealed that brute force has a very poor processing time, while sorting and neighbourhood frequency does not have issues with processing time. Conclusion: The research revealed that existing anagram orthographic similarity measures are not suitable for character position verification and evaluation of syllabic complexity which are essential measures of working memory capacity. Keywords: Cognition, orthography, similarity measures, bi-anagram, permutation.

All co-authors agreed to have their names listed as authors.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2018 The Authors. Journal of Research and Reviews in Science – JRRS, A Publication of Lagos State University

JRRS https:sciencejournal.lasucomputerscience.com

173

Page 175

1. INTRODUCTION like aphasia disease through anagram autographic similarity measure. [9],[10] The word permutation has to do with the • It is necessary to improvise for a rearrangement of characters in a word to form cognitive career counsellor. This is due other . Several new words can be formed to the fact that, cognitive test reveals from a single word through word permutation. working memory capacity, and can in The rearrangement of the characters of a word turn be used to predict student’s ability to form another word of the same length n-gram to withstand some courses of study. but different position is called anagram. This is One of the steps to this is performance relevant for user’s cognitive analysis [8], evaluation of anagram autographic analysis of user’s memory capacity, analysis of similarity measure, which will reveal the user’s word fluency through anagram solving lapses of existing measures for ability. enhancement. [8] Verifying the memory stability of individual is • Several applications had swept mundane at certain levels. In cognitive deficit students attention from reading, which people due to Alzheimer’s disease, during the is a defect factor of student’s reading recovery process or medical monitoring of fluently. Anagram solving task is fun fill patients, it is important that patients are and also educative in terms of word at observed psychologically from time to time. sight and reading fluently for students. Anagramming is a very good measure of [5], [7], [11] recovery level in this case. The psychological status can further be measured through the 2. MATERIAL AND METHODS kind of anagram that can be solved by patients. The authors in [1] researched on the effects of [9,10]Ability of patients to generate hard or syllabic structure and grapheme frequency of average or simple anagram can quantify their target words on an anagram solving task. He level of recovery in terms of cognitive ability. also checked for presumable differences on the While choosing career during teenage age, solving performance of average and below cognitive ability can depict memory smartness average readers. This was achieved by of individuals. Challenging courses like medical measuring time spent and user’s number of sciences, engineering, law e.t.c. requires moves for solving different non-sense candidates with high cognitive ability. This is to anagrams. Data analysis revealed that enhance and ease student’s study of a course. anagram solving time is affected by the syllabic Anagram can serve as a good measure of structure of target words. The main effect of candidate’s intellectual capacity. Thus, it can be syllabic complexity indicates that both groups used as one of the measures for taking decision i.e. average and below average readers were during career counselling of teenagers. It gives equally affected. The effect of syllabic a true picture of working memory capacity of complexity was also noticed in the reading students. Also, Anagram solving can be used to fluency measure, where the above average enhance student’s mastery ability. group out perform the below average group. Anagrams also serve as a means of data The two experiment groups achieve lower time encryption which is applicable to network scores at the simple syllabic structure. The security.[5,7,11] syllabic structures used in the research are: consonant-vowel(CV), consonant-consonant- This research explored anagram orthographic vowel(CCV), consonant-vowel-vowel(CVV) similarity measures, by considering their and consonant-vowel-consonant(CVC). An strengths and draw backs. This is to determine anagram solving task can be conceptualized as the necessary enhancement needed by these a lexical access task. [2] measures. The authors in [3] investigated lexical decision Anagram orthographic similarity measures task, to be able to deduce that “poor motion” were analysed, and thus serves as analysis detectors should encode letter position less measure for anagrams solved by users for accurately than good motion detectorss. He various purposes. Through these measures thus, inferred that “poor motion detector” should deductions can be drawn on users based on be more likely to unscramble briefly presented their anagram solving ability. anagrams and respond to them as if they were words. 1.1 Motivation The authors in [4] introduced the use of • It is necessary to devise a standard anagram for resisting evasion by polymorphic means of monitoring recovery rate and mimicry attacks, randomization strategy was level of patients with cognition problem used to achieve this.

LASU Journal of Research and Review in Science Page 176

One of the tools used in psychology to provides different modes of operation: investigate cognitive processes is anagrams. interactive and automatic. The interactive mode This research applied rasch scaling to a sample allows a researcher to provide a hard generated of five-letters anagrams to examine whether a list of test strings or words for evaluation of single undimensional scale based on syllable possible anagram. The list can be entered number can be usefully applied. Rasch analysis either interactively or through an input file. The was used to establish a participant’s ability to automatic mode generates a specific number of solve anagrams, and also to establish the fixed length test string pseudo randomly. In relative difficulty of each anagram. Rasch’s either case, the test string is converted to a theory suggests that the probability of getting sorted order, and each possible sub-string is an anagram correct is caused by the difference considered. All possible anagrams are in a person’s ability and the anagram’s difficulty. identified, and the frequency information It was deduced from the research that syllable for all orthographically identical word forms is number is a major factor in determining summed and printed. The research did not anagram difficulty. The more syllables the consider bi-gram frequency in anagrams. target word contain, the harder it was to solve and the higher the rasch score. Thus, syllable The processes involved in working through number is a compounding factor for other problems to solve anagram was reported in [9]. variables that contribute to anagram difficulty. The authors looked into the role of effort in This research made a useful contribution to remembering anagrams and their solutions. measurement models of human cognitive Firstly, experiment was conducted to compare problem solving [5]. the effect on memory of copying word, typing them as solution for easy/hard anagrams, and The authors in [6] used high-density (64 solving anagrams. Secondly, memory for channel) event related potential(ERP) anagram using frequency judgement was combined with di-pole source analysis to compared. It was deduced that memory for explore the neural bases of unconscious error difficult anagram is not better than memory for detection in an anagram solution task. ERP easy one. studies indicates error negativity (Ne/ERN) and error positivity(Pe), they are used for error Sandra et. al. in [10] researched on a new monitoring and error detection. Responses measure, the Northwestern Anagram Test were collected for unconscious error (NAT), which was developed to test accuracy of detection(UED), No error(NE) and detection word order (syntax) in sentence production in error(DE), the response time was also recorded patients with speech production, word i.e. the time from when anagram appear and comprehension and/or word-finding difficulties, response. The research analysed the ERPs or reduced working memory capacity. The elicited by the Chinese anagrams, and deduced anagram method requires the assembly of that amplitude of N2 i.e. negative ERP elicited individual word cards presented in scrambled by the UED response was greater than those of order into meaningful sentences. A related test NE and DE responses. Late positive for aphasia, the verb and sentence test component elicited by the DE response was (VAST17), includes an anagram task; however, greater than those of the UED and NE its major purpose is to evaluate verb and verb responses. argument structure. Thus, the stimulus cards group words to form complete sentence Anagram priming effect can be found in lexical constituents (e.g, complete noun phrases: The decision task, with non-adjacent transposed boy / is chasing / the dog.) on the other hand, letters. Anagram relationship within or the VAST examines only active and passive between stimuli, constitute a form of sentence structures. The NAT provides a orthographic similarity whose effect on reading separate stimulus card for each word of the performance is detectable. orthographic target structures such that the patient has few similarity on anagram can be detected using clues as to permissible word combinations. In neibourhood frequency effect, while anagram addition, like the NAVS, the NAT examines relationship acts as orthographic similarity several canonical forms. The research report factor [7]. results obtained with the NAT from patients with primary progressive aphasia (PPA) and their Researchers in [7] worked on effect of anagram cognitively intact controls. Northwestern relationship between lexical units, prime and Anagram Test performance was compared with target stimuli, the research deduced that mental performance on the sentence production lexicon is activated by position-free letter primary test(SPPT) of the NAVS and was codes, together with other units that encode the compared with measures of aphasia severity, order information. naming, single word comprehension, oral Robert and Debra in [8] worked on anagram apraxia, and object recognition to demonstrate software for cognitive research, the software its specificity for syntax. It was hypothesized LASU Journal of Research and Review in Science Page 177

that the NAT would correlate with measures of of solution word, word imagery, correctness of syntax but not with measures of naming, word word familiarity, age of acquisition, comprehension, or motor speech production. meaningfulness of vowels, staring letter, some Furthermore, it was predicted that performance bigram frequency measures. Further research would be influenced by grammatical complexity revealed that the most important measure for [10]. difficulty is bigram frequency. The canonical well known model of word Anagram tasks are frequently used in recognition, which stem from Interactive behavioural research to investigate a wide Activation Model of Word and Letter array of cognitive phenomena. Most Perception, use slot-filler representations for prominently, they are used to study the letters, and are thus poorly suited to capture cognitive stages involved in problem solving, letter permutation effects. Henin et. al. in [11] specifically insight [8]. proposed a solution that takes up a prominent feature of the early interactive activation Researchers in [1] affirmed that anagram models: units corresponding to structures at solution tasks have been frequently used to multiple spatial (and temporal) scales (e.g. assess word recognition processes. Also , bigram, word). Unlike in the anagram solution ability is closely related to canonical models, the units in this model detect reading. Reading is an innate skill, it is one of structure independently of spatial location. the most crucial cognitive skill that support They also activate partially in response to school based learning. This is because it is a partial similarity. As a result, the model is order- lengthy process that requires mastering a large sensitive without being order-rigid. If the set of strategies. network stabilizes on a pattern which is not a Anagram solving task are considered to tap solution, the model identifies a symbol ordering high level cognitive abilities [15]. The anagram consistent with the current state and uses this paradigm had been used in cognitive research symbol ordering to reset the state for another to assess word recognition process.[3]. In episode of stabilization. This model correctly anagram solving, bigram frequency is an predicts the positive correlations between high important feature, anagrams derived from bigram frequency in the target and solution target words with high bigram frequency are time, and between low bigram frequency in the easier and vise versa [16]. stimulus and solution time. However, because the constraint satisfaction process is not Some cognitive tasks like reading, spelling, constrained by lexical information, the model making lexical decision, solving anagram task does not seem well-suited to modelling normal e.t.c. requires knowledge of orthographic reading. The model, NGRAMSWELL, was structure of English. Examples of orthographic proposed by this researcher [11]. structure are syllabification, orthographic neighbourhood size, and bigram frequency Authors in [12] carried out experiment to show [12]. that type-based(bi-gram of a particular word) bi- gram frequency is a better predictor of the The authors in [17] discovered that computer difficulty of anagram solution than is token- based anagram generators do not provide based( all occurrences of bigram in a corpus) controls necessary for psychological research. frequency. The research used brute force algorithm to automatically process words in a user defined A series of orthographic measures for source and output, for each word all psycholinguistic research was presented in possible anagrams that exist are defined by the [13]. Orthographic measure factors are word same source vocabulary. length, word-form frequency, lemma frequency, neighbourhood density and frequency, Experiment was conducted to find out cognitive transposition neighbours, addition and deletion flexibility, this was performed using anagram neighbours are essential. word puzzles. It was observed that contrast in performance between WAKE, REM(dreaming) The research conducted in [14] showed that a and NREM suggests that cognitive processing new bi-gram frequency measure called top differs in the three states. It can be inferred from rank, is an important predictor of anagram this research that anagram is a true measure of difficulty, previous researchers opined that the cognitive abilty [18]. type count are better than the token measures at predicting anagram difficulty. Previous Researches on anagram had explored different researches suggested many variables that methods for detecting orthographic similarity influence the difficulty of anagram i.e. features between anagrams. Methods like Brute force Neighbourhood frequency i.e. counting and [17], Sorting [8], Bubble sort [11], histogram [1, 7, 9].

LASU Journal of Research and Review in Science Page 178

Table 1. Previous Approaches to Orthographic Analysis of Anagram Task

S/N Authors Method Pro Cons 1 Menalaos and Syllabic structure It unveiled how: The correlation Chris [1] and grapheme -Syllabic structure between the frequency of affects anagram recognition of target words on solving time anagram containing anagram solving -Syllabic complexity infrequent task has effect on reading grapheme and fluency measure those with frequent phoneme-to- grapheme mapping is not explored

2 Norvick and Position-sensitive It predicts the It is restricted to five Shennem [12] type-based bi- difficulty of anagram letters word gram frequency solution compared to token based numbers 3 Sergio et al. [4] Applied rasch To establish Syllable number is scaling to syllabic individual’s ability to restricted to five letters. structure solve anagram and the relative difficulty of anagram 4 Hua-Zhan Yin Syllabic solution It depicts Orthography is restricted et al. [6] of anagram unconscious and to Chinese language. conscious error detection in anagram solution task 5 Couriear and Orthographic Established Research was Lequexx [7] neighbourhood relationship between conducted using size lexical units, prime analysis tool. There was and target stimuli no standard software for implementation 6 Mary et. al. [9] Syllabic structure -Syllabic structure Research was and orthographic was used to compare conducted using neighbourhood the effect of memory analysis tool. There was frequency on copying words no standard software for -Neighbourhood implementation frequency was used to compare memory for anagram 7 Sandra et. al. Exploration of It is used to evaluate Anagram test was [10] syllabic structure the cognitive ability conducted orally of patients with primary progressive aphasia 8 Robert et. al. Syllabic structure It enables direct -Bigram frequency [8] using sorting control of calculation was not anagram psycholinguistic directly incorporated in detection features that may the software technique influence the -Orthographic cognitive process in neighbourhood anagram solution frequency is not explored

LASU Journal of Research and Review in Science Page 179

9 Henin et. al. The orthographic It serves as better Bubble sort makes [11] structure used is predictor of anagram execution using N-Gram N-GRAM WELL, detection. This is slower with bubble sort done by using bubble as a measure of sort to find the anagram distance detection transforming a supplied anagram answer to target word. -It is position free

11 Matthew [18] Syllabic structure Anagram was used Anagram task was rearrangement to predict cognitive restricted to five letters. flexibility

12 Thimothy and Anagram Words can be Brute force has a poor Axel [17] permutation processed running time algorithm(Brute automatically in user force) using defined source syllabic structure vocabulary and rearrangement output.

Table 1 shows that researchers used one or private boolean areAnagrams(String first, more orthographic structure for anagram String second) { analysis. The following are draw backs of previous researches. return areAnagramsRec("", first, second); }

• Restriction of anagram letters to five [4, 12] private boolean areAnagramsRec(String • Use of analysis tool, no standard software soFar, String remaining, String target) { was developed [7, 9]. if (remaining.length() == 0) • Oral conduction of anagram test, no standard software was developed [10]. { return soFar.equals(target); } • The existing cognitive software does not for (int i = 0; i < remaining.length(); i++) have Bigram orthographic structure incorporated into it. It uses sorting { String whatsLeft = detection technique i.e. no syllabic remaining.substring(0, structure relationship detection [8] i)+remaining.substring(i+1); • Bi-gram frequency with bubble sort anagram detection without consideration if (areAnagramsRec(soFar + of position of characters [11] remaining.charAt(i), whatsLeft, target)) • Poor anagram detection processing time return true; } return false; } [11, 17]. 3.1.2 Sorting: Two strings are anagrams of 3.1 Existing Anagram Orthographic each other if they are equal, when their letters Similarity Measures are sorted. Thus, the presence of anagram can The presence of anagram within two strings can be detected by sorting the characters in each be detected by using any of the following string, and test if the characters in both strings methods: are equal. The method for anagram detection for sorting (ADTS) is given as follows [19]: 3.1.1 Brute force: This has to do with listing all permutations of the first string, and check if the private boolean areAnagrams(String first, second string is equal to any of the String second) permutations of the first. This gives a very poor { char[] complexity n!. The algorithm for brute force is one=Arrays.sort(first.toCharArray()); given as follows [19]:

Page 180

char[] two = If the strings are anagram the result will be Arrays.sort(second.toCharArray()); return Arrays.equals(one, two); } one, because the number of counts of all ADTS( x , y )= 1 − S , S −  S S ...... 1characters will be zero. xy ij 3.1.4 Histogram: This works by building where Si and Sj are characters of sorted list frequency histogram of characters in each of strings X and Y. string, and check whether both strings are the same. This method is not explored because it’s |Sx| and |Sy| represent length of sorted list execution time is similar to counting method. X and Y The methods listed above i.e. brute force, SS finds out the presence of a ij sorting, neighbourhood frequency: counting and histogramming were explained above. character in both strings, and add 1 to sum Their complexities are O(n! ), O(nlogn), O(n2) for each presence. The sum of the ,O(n). [19] Thus, brute force runs with the intersection must be equal to the length of poorest running time, followed by counting, X as well as Y for anagram to hold. sorting and then histogramming. The research 3.1.3 Counting: It has to do with counting how is thus considering sorting and counting, these many times each character occur in each string, two are preferred because histogram is faster and confirm that each string has the same but it has trade off for space, while brute force number of character as the other. The method is too poor in terms of running time. for anagram detection for counting (ADTC) is given as follows [19]: The techniques analyse anagram in words by private boolean areAnagrams(String first, considering the occurrence of characters String second) { without taking cognisance of position of for (char ch = 'a'; ch <= 'z'; ch++) { characters. For example, to analyse the strings if (numCopiesOf(ch, first) != “ALLERGY” and “ALLERGY” numCopiesOf(ch, second)) anagrammatically, the existing methods test for { return false; } return true; } the occurrence of all the characters in the first private int numCopiesOf(char ch, String and in the second word without considering str) { their positions. The two methods adopted in this int result = 0; work are ADTS and ADTC. Thus, these for (int i = 0; i < str.length(); i++) { mesures analyses two equal strings as if (str.charAt(i) == ch) result++; } anagram. return result; } 3. RESULTS AND DISCUSSION

The algorithm above is interpreted into the 3.1 Implementation and Results following mathematical equation for easy evaluation. The methods were implemented using java nm, ADTC( x , y )= 1 − C − C programing ...... 2 language, the back end was ijij==0, 0 implemented using SQLite studio. Eigthy seven

C – represents the number of count of unique words were autographically examined. each character i.e. frequency Correct anagrams were extracted from each

word and populated in the anagram

which serves as the dataset. Experiment was

Page 181

serial number 4. All the methods evaluates it to conducted many times with different dataset. zero. Meaning that it’s not an anagram 3. Text equivalent to Pattern: Exactly the Experiments were carried out on the dataset same string was supplied as anagram on serial number 6. ADTS and ADTC evaluate to 1 using the following procedures: because it sees them as anagram, these methods does not consider position of strings, Figure 1 is the interface for analysing a given rather it considers occurrence of strings. 4. Absence of Pattern in the string and supplied string by user for anagram dictionary(user defined vocabulary): A wrong string was supplied as anagram on serial relationship. A word is selected from the number 8, ADTS and ADTC considers it as anagram because it doesn’t do correct word database, and the user is to supply one to five verification before returning output. anagrams depending on the number of 5. Absence of character entailment between Text and Pattern: A wrong word was anagrams that can be generated from a word. 6. supplied as anagram on serial number 10. The methods returned 0. The submit button triggers anagram test for Table 2 shows the anagram status and cognitive ability. Anagram test is performed processing time of each text and pattern. The processing time of the dataset for each method using sorting and counting methods, these was averaged, this indicates that ADTS is the faster, followed by ADTC, just as revealed by methods were used to compare the the method’s complexity. orthographic structure of strings using syllabic Experiments on the measures revealed that, tested measures only test for anagram through frequency check. The methods considers the character entailment verification. Moreover, in anagram, position of characters is very occurrence of all characters in sting1 and important, these measures doesn’t do position verification, which might amount to analysing string2, without considering the position of same words presented as given and target characters which is key in anagram detection. words as anagram which is the case in Table 2, ADTS and ADTC takes care of anagram of the No. 6. Also, apart from conducting orthographic same length, they both returns value 1 in this anagram similarity measure, these measures case, indicating that anagram is detected. If the can not test for syllabic relationship between text and pattern are equal they evaluate to 1, anagram strings, which actually reveal the because they check for character occurrence. distance between two anagrams, and thus In ADSM, it’s value is either 0 or 1, thus, if text unveil their complexity i.e. simple, average or and pattern has unequal length and equal hard anagram. strings, it evaluates to zero, and strings with same character composition but different position evaluates to 1. ADTS and ADTC does not have any function incorporated into it, for detection of wrong anagram supplied by students i.e. orthographic analysis. Table 2 depicts the result of anagram test of strings (sample words) using different conditions, this is to show the reaction of these methods to the stipulated conditions. The conditions are explained below:

1. Text longer than Pattern: A string of shorter length was supplied as anagram on Figure 1: Main View of Orthographic serial number 2 i.e. pattern. All the methods evaluates it to zero. Meaning that it’s not an Anagram Analysis Interface anagram 2. Text shorter than Pattern: A string of Longer length was supplied as anagram on

Page 182

Table 2: Anagram Methods Running Time Analysis Table for Sample Words

S/NO. TEST WORD/ USER’S ADTS ADTC Anagram

ANSWER Anagram value;

value; Processing Time

Processing (*10^-5)secs

Time

(*10^-5)secs

1 Auctioned/Cautioned 1; 2.3 1; 65.0

2 Auctioned/educatio 0; 2.3 0; 65.0

3 Allergy/Regally 1; 1.9 1; 07.0

4 Allergy/Galllery 0; 1.9 0; 07.0

5 Antler/Rental 1; 2.3 1; 7.5

6 Antler/Antler 1; 2.3 1; 7.5

7 Ales/Leas 1; 1.5 1; 3.3

8 Ales/Laes 1; 1.5 1; 3.3

9 Assert/Asters 1; 1.8 1; 4.4

10 Asserts/Strong 0; 1.8 0; 4.4

Mean 1.78 17.44

Page 183

4. CONCLUSION 7. Lequex, C.P.a., anagram effect in visual word recognition. 2004, https://archives- Previous approaches to orthographic similarity of onverts.fr/hat0042984. anagrams were based on Brute force, Sorting, 8. Robert D.V, Y.K.G., Debra A.T Anagram Orthographic neighborhood frequency. The software for cognitive research that enables experimented measures were picked based on their specification of Pscholinguistic variables. running time complexities and memory space. Data sets Behaviour Research Methods, Instruments and were tested on the measures for various conditions to Computers, reveal the behaviour of the measures under different http://dx.doi.org/10.3758/BF03195587, 2006. circumstances. User defined and 38(2). orthographic parameters was used for orthographic 9. Mary A F, H.J.F., Alice W., Leslie R. , Anagram verification. Experiment revealed that the measures has solving: Does effort have an effect. memory and the capacity to test for orthographic similarity of cognition, 1989. 17(6): p. 755-758. anagram through character entailment verification only. 10. Sandra W., M.-M.M., Christina W, Alfred R, The drawback thus, lies in lack of character position Emily J.R, Cynthia K.T. , The Northerwestern verification and syllabic relationship test which are very Test: Measuring sentence production in vital while testing user’s working memory capacity. primary progressive Aphasia. American Journal ACKNOWLEDGEMENTS of Alzhheinder’s disease and other Dementias, 2009. We want to use this medium to acknowledge the efforts 11. Henin, J., et al. Extraordinary natural ability: of the laboratory technicians of Computer Science Anagram solution as an extension of normal Department, Lagos State University for their technical reading ability. in Proceedings of the 31st support in the course of this research, i.e. Mr Mayowa Annual Meeting of the Cognitive Science and Mr Bamiro. We also want to commend the effort of Society. 2009. Mr Ekundayo a Software analyst for his technical 12. S.J., N.L.R.a.S., Type based bigram support during the program implementation phase of frequencies for five letter words. Behaviour this research. This study is self-sponsored. Research Methods, Instruments and Computers, AUTHORS’ CONTRIBUTIONS http://dx.doi.org/10.3758/BF03195587, 2004. 36(3): p. 397-401. Raji-Lawal Hanat Y. designed the study, performed the 13. Maria Ktori, W.J.B., Van Heuven, Nicola J statistical analysis, wrote the protocol, and wrote the Ptchford, Greek Lex: A lexical database of first draft of the manuscript. Akinwale Adio T. modern greek. Behaviour Research Methods, Folorunsho O., and Mustapha Amidu O. conducted 2008. 40(3): p. 773-783. literature searches for the study. All authors read and 14. David Knight, S.J.M., Type and token bigram approved the final manuscript.” frequencies for two trough nine letter words and the prediction of anagram difficulty. Behaviour Research Journal, 2011. 43: p. 491-498. 15. Mozer, G.a., The interplay of symbolic and sub- REFERENCES symbolic processes in anagram problem solving. Advances in neural information 1. Menelaos E.S and Chris T.P, . Linguistic effect processing system 2001. 13: p. 17-23. on anagram solution: The case of transparent 16. Johnson, G., Effect of solution word attributes language. world class journal of education, on anagram difficulty. A regression analysis 2013. 3( 4): p. 41-51. quarterly journal of experimental psychology, 2. Fink, T.E. and R.W. Weisberg, The use of 1978. 30: p. 57-70. phonemic information to solve anagrams. 17. Monteiro, T.R.J.a.A., Generating anagrams Memory & Cognition, 1981. 9(4): p. 402-410. from multiple core strings employing user 3. Cornelissen, P., et al., Coherent motion defined vocabularies and orthographic detection and letter position encoding. Vision parameters. behaviour Research Methods, Research, 1998. 38(14): p. 2181-2191. Instruments and Computers, 2003. 35(1): p. 4. Sergio P., A.O., Juan E.T., Pedro P., 129-135. Randomized anagram revisited. Cosec 18. Matthew P. Walker, C.L., J. Allan Hobson, Laboratory, Elsevier, 2014. Robert Stick Gold, Cognitive flexibility across 5. Adams, J.W., et al., The role of syllables in sleep-wake cycle: REM-Sleep enhancement of anagram solution: A Rasch analysis. The anagram solving. Elsevier Journal of cognitive Journal of general psychology, 2011. 138(2): p. brain research, 2002. 14: p. 317-324. 94-109. 19. University, S., Anagram detection 6. Hua-Zhan Yin, J.Y., Wei Li, Jiang Qui, Ling Yu web.stanford.edu/class/cs9/Sample- Chen Neural bases of unconscious error Probs/Anagram.pdf detection in a Chinese anagram solution task: Evidence from ERP study. Journal Pone 0154379, 2016.