Main Content

transform

Transform documents into lower-dimensional space

Description

dscores = transform(lsaMdl,documents) transforms documents into the semantic space of the latent semantic analysis (LSA) model lsaMdl.

example

dscores = transform(lsaMdl,bag) transforms documents represented by the bag-of-words or bag-of-n-grams model bag into the semantic space of the LSA model lsaMdl.

dscores = transform(lsaMdl,counts) transforms documents represented by the matrix of word counts into the semantic space of the LSA model lsaMdl.

dscores = transform(ldaMdl,documents) transforms documents into the latent Dirichlet allocation (LDA) topic probability space of LDA model ldaMdl. The rows of dscores are the topic mixture representations of the documents.

example

dscores = transform(ldaMdl,bag) transforms documents represented by the bag-of-words or bag-of-n-grams model bag into the LDA topic probability space of LDA model ldaMdl.

dscores = transform(ldaMdl,counts) transforms documents represented by the matrix of word counts into the LDA topic probability space of LDA model ldaMdl.

example

dscores = transform(___,Name,Value) specifies additional options using one or more name-value pair arguments. These name-value pairs only apply if the input model is an ldaModel object.

Examples

collapse all

Load the example data. The file sonnetsPreprocessed.txt contains preprocessed versions of Shakespeare's sonnets. The file contains one sonnet per line, with words separated by a space. Extract the text from sonnetsPreprocessed.txt, split the text into documents at newline characters, and then tokenize the documents.

filename = "sonnetsPreprocessed.txt";
str = extractFileText(filename);
textData = split(str,newline);
documents = tokenizedDocument(textData);

Create a bag-of-words model using bagOfWords.

bag = bagOfWords(documents)
bag = 
  bagOfWords with properties:

          Counts: [154x3092 double]
      Vocabulary: ["fairest"    "creatures"    "desire"    "increase"    "thereby"    "beautys"    "rose"    "might"    "never"    "die"    "riper"    "time"    "decease"    "tender"    "heir"    "bear"    "memory"    "thou"    ...    ] (1x3092 string)
        NumWords: 3092
    NumDocuments: 154

Fit an LSA model with 20 components.

numCompnents = 20;
mdl = fitlsa(bag,numCompnents)
mdl = 
  lsaModel with properties:

              NumComponents: 20
           ComponentWeights: [2.7866e+03 515.5889 443.6428 316.4191 295.4065 261.8927 226.1649 186.2160 170.6413 156.6033 151.5275 146.2553 141.6741 135.5318 134.1694 128.9931 124.2382 122.2931 116.5035 116.2590]
             DocumentScores: [154x20 double]
                 WordScores: [3092x20 double]
                 Vocabulary: ["fairest"    "creatures"    "desire"    "increase"    "thereby"    "beautys"    "rose"    "might"    "never"    "die"    "riper"    "time"    "decease"    "tender"    "heir"    "bear"    "memory"    ...    ] (1x3092 string)
    FeatureStrengthExponent: 2

Use transform to transform the first 10 documents into the semantic space of the LSA model.

dscores = transform(mdl,documents(1:10))
dscores = 10×20

    5.6059   -1.8559    0.9286   -0.7086   -0.4652    0.8340    0.6751   -0.0611   -0.2268    1.9320   -0.7289   -1.0864    0.7131   -0.0571   -0.3401    0.0940   -0.4406    1.7507   -1.1534    0.1785
    7.3069   -2.3578    1.8359   -2.3442   -1.5776    2.0310    0.7948   -1.3411    1.1700    1.8839    0.0883    0.4734   -1.1244    0.6795    1.3585   -0.0247    0.3627   -0.5414   -0.0272   -0.0114
    7.1056   -2.3508   -2.8837   -1.0688   -0.3462    0.6962    0.0334    0.0472   -0.4916    0.6496   -1.1959   -1.0171   -0.4020    1.2953   -0.4583    0.5984   -0.3890    1.1780    0.6413    0.6575
    8.6292   -3.0471   -0.8512   -0.4356   -0.3055   -0.4671   -1.4219    0.8454    0.8270    0.4122    2.2082   -1.1770    1.7775   -2.2344   -2.7813    1.4979    0.7486   -2.0593    0.6376    1.0721
    1.0434    1.7490    0.8703   -2.2315   -1.1221   -0.2848   -2.0522    0.6975   -1.7191   -0.2852    0.8879    0.9950   -0.5555    0.8842   -0.0360    1.0050    0.4158    0.5061    0.9602    0.4672
    6.8358   -2.0806   -3.3798   -1.0452   -0.2075   -2.0970   -0.4477   -0.2080   -0.9532    1.6203    0.6653    0.0036    1.0825    0.6396   -0.2154   -0.0794    0.7108    1.8007   -4.0326   -0.3872
    2.3847    0.3923   -0.4323   -1.5340    0.4023    1.0396   -1.0326   -0.3776   -0.2101   -1.0944   -0.7513   -0.2894    0.4303    0.1864    0.4922    0.4844    0.5191   -0.2378    0.9528    0.4817
    3.7925   -0.3941   -4.4610   -0.4930    0.4651   -0.3404   -0.5493   -0.1470   -0.5065    0.2566    0.3394   -1.1529   -0.0391   -0.8800   -0.4712    0.9672    0.5457   -0.3639   -0.3085    0.5637
    4.6522    0.7188   -1.1787   -0.8996    0.3360   -0.4531   -0.1935   -0.3328    0.8640   -1.6679   -0.8056   -2.1993    0.1808    0.0163   -0.9520   -0.8982    0.6603    3.6451    1.2412    1.9621
    8.8218   -0.8168   -2.5101    1.1197   -0.8673    1.2336    0.0768   -0.1943    0.7629   -0.1222    0.3786    1.1611    0.2326    0.3415   -0.3327   -0.3792    1.7554    0.2526   -2.1574   -0.0193

To reproduce the results in this example, set rng to 'default'.

rng('default')

Load the example data. The file sonnetsPreprocessed.txt contains preprocessed versions of Shakespeare's sonnets. The file contains one sonnet per line, with words separated by a space. Extract the text from sonnetsPreprocessed.txt, split the text into documents at newline characters, and then tokenize the documents.

filename = "sonnetsPreprocessed.txt";
str = extractFileText(filename);
textData = split(str,newline);
documents = tokenizedDocument(textData);

Create a bag-of-words model using bagOfWords.

bag = bagOfWords(documents)
bag = 
  bagOfWords with properties:

          Counts: [154x3092 double]
      Vocabulary: ["fairest"    "creatures"    "desire"    "increase"    "thereby"    "beautys"    "rose"    "might"    "never"    "die"    "riper"    "time"    "decease"    "tender"    "heir"    "bear"    "memory"    "thou"    ...    ] (1x3092 string)
        NumWords: 3092
    NumDocuments: 154

Fit an LDA model with five topics.

numTopics = 5;
mdl = fitlda(bag,numTopics)
Initial topic assignments sampled in 0.064998 seconds.
=====================================================================================
| Iteration  |  Time per  |  Relative  |  Training  |     Topic     |     Topic     |
|            | iteration  | change in  | perplexity | concentration | concentration |
|            | (seconds)  |   log(L)   |            |               |   iterations  |
=====================================================================================
|          0 |       0.01 |            |  1.212e+03 |         1.250 |             0 |
|          1 |       0.01 | 1.2300e-02 |  1.112e+03 |         1.250 |             0 |
|          2 |       0.01 | 1.3254e-03 |  1.102e+03 |         1.250 |             0 |
|          3 |       0.01 | 2.9402e-05 |  1.102e+03 |         1.250 |             0 |
=====================================================================================
mdl = 
  ldaModel with properties:

                     NumTopics: 5
             WordConcentration: 1
            TopicConcentration: 1.2500
      CorpusTopicProbabilities: [0.2000 0.2000 0.2000 0.2000 0.2000]
    DocumentTopicProbabilities: [154x5 double]
        TopicWordProbabilities: [3092x5 double]
                    Vocabulary: ["fairest"    "creatures"    "desire"    "increase"    "thereby"    "beautys"    "rose"    "might"    "never"    "die"    "riper"    "time"    "decease"    "tender"    "heir"    "bear"    ...    ] (1x3092 string)
                    TopicOrder: 'initial-fit-probability'
                       FitInfo: [1x1 struct]

Use transform to transform the documents into a vector of topic probabilities. You can visualize these mixtures using stacked bar charts. View the topic mixtures of the first 10 documents.

topicMixtures = transform(mdl,documents(1:10));
figure
barh(topicMixtures,'stacked')
xlim([0 1])
title("Topic Mixtures")
xlabel("Topic Probability")
ylabel("Document")
legend("Topic " + string(1:numTopics),'Location','northeastoutside')

Figure contains an axes object. The axes object with title Topic Mixtures, xlabel Topic Probability, ylabel Document contains 5 objects of type bar. These objects represent Topic 1, Topic 2, Topic 3, Topic 4, Topic 5.

Load the example data. sonnetsCounts.mat contains a matrix of word counts and a corresponding vocabulary of preprocessed versions of Shakespeare's sonnets.

load sonnetsCounts.mat
size(counts)
ans = 1×2

         154        3092

Fit an LDA model with 20 topics. To reproduce the results in this example, set rng to 'default'.

rng('default')
numTopics = 20;
mdl = fitlda(counts,numTopics)
Initial topic assignments sampled in 0.072024 seconds.
=====================================================================================
| Iteration  |  Time per  |  Relative  |  Training  |     Topic     |     Topic     |
|            | iteration  | change in  | perplexity | concentration | concentration |
|            | (seconds)  |   log(L)   |            |               |   iterations  |
=====================================================================================
|          0 |       0.00 |            |  1.159e+03 |         5.000 |             0 |
|          1 |       0.02 | 5.4884e-02 |  8.028e+02 |         5.000 |             0 |
|          2 |       0.03 | 4.7400e-03 |  7.778e+02 |         5.000 |             0 |
|          3 |       0.02 | 3.4597e-03 |  7.602e+02 |         5.000 |             0 |
|          4 |       0.03 | 3.4662e-03 |  7.430e+02 |         5.000 |             0 |
|          5 |       0.02 | 2.9259e-03 |  7.288e+02 |         5.000 |             0 |
|          6 |       0.02 | 6.4180e-05 |  7.291e+02 |         5.000 |             0 |
=====================================================================================
mdl = 
  ldaModel with properties:

                     NumTopics: 20
             WordConcentration: 1
            TopicConcentration: 5
      CorpusTopicProbabilities: [0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500 0.0500]
    DocumentTopicProbabilities: [154x20 double]
        TopicWordProbabilities: [3092x20 double]
                    Vocabulary: ["1"    "2"    "3"    "4"    "5"    "6"    "7"    "8"    "9"    "10"    "11"    "12"    "13"    "14"    "15"    "16"    "17"    "18"    "19"    "20"    "21"    "22"    "23"    "24"    "25"    ...    ] (1x3092 string)
                    TopicOrder: 'initial-fit-probability'
                       FitInfo: [1x1 struct]

Use transform to transform the documents into a vector of topic probabilities.

topicMixtures = transform(mdl,counts(1:10,:))
topicMixtures = 10×20

    0.0167    0.0035    0.1645    0.0977    0.0433    0.0833    0.0987    0.0033    0.0299    0.0234    0.0033    0.0345    0.0235    0.0958    0.0667    0.0167    0.0300    0.0519    0.0833    0.0300
    0.0711    0.0544    0.0116    0.0044    0.0033    0.0033    0.0431    0.0053    0.0145    0.0421    0.0971    0.0033    0.0040    0.1632    0.1784    0.0937    0.0683    0.0398    0.0954    0.0037
    0.0293    0.0482    0.1078    0.0322    0.0036    0.0036    0.0464    0.0036    0.0064    0.0612    0.0036    0.0176    0.0036    0.0464    0.0906    0.1169    0.0888    0.1115    0.1180    0.0607
    0.0055    0.0962    0.2403    0.0033    0.0296    0.1613    0.0164    0.0955    0.0163    0.0045    0.0172    0.0033    0.0415    0.0404    0.0342    0.0176    0.0417    0.0642    0.0033    0.0676
    0.0341    0.0224    0.0341    0.0645    0.0948    0.0038    0.0189    0.1099    0.0187    0.0560    0.1045    0.0356    0.0668    0.1196    0.0038    0.0931    0.0493    0.0038    0.0038    0.0626
    0.0445    0.0035    0.1167    0.0034    0.0446    0.0583    0.1268    0.0169    0.0034    0.1135    0.0034    0.0034    0.0047    0.0993    0.0909    0.0582    0.0308    0.0887    0.0856    0.0034
    0.1720    0.0764    0.0090    0.0180    0.0325    0.1213    0.0036    0.0036    0.0505    0.0472    0.0348    0.0477    0.0039    0.0038    0.0122    0.0041    0.0036    0.1605    0.1487    0.0465
    0.0043    0.0033    0.1248    0.0033    0.0299    0.0033    0.0690    0.1699    0.0695    0.0982    0.0033    0.0039    0.0620    0.0833    0.0040    0.0700    0.0033    0.1479    0.0033    0.0433
    0.0412    0.0387    0.0555    0.0165    0.0166    0.0433    0.0033    0.0038    0.0048    0.0033    0.0473    0.0474    0.1290    0.1107    0.0089    0.0112    0.0167    0.1555    0.2423    0.0040
    0.0362    0.0035    0.1117    0.0304    0.0034    0.1248    0.0439    0.0340    0.0168    0.0714    0.0034    0.0214    0.0056    0.0449    0.1438    0.0036    0.0290    0.1437    0.0980    0.0304

Input Arguments

collapse all

Input LSA model, specified as an lsaModel object.

Input LDA model, specified as an ldaModel object.

Input documents, specified as a tokenizedDocument array, a string array of words, or a cell array of character vectors. If documents is a tokenizedDocument, then it must be a column vector. If documents is a string array or a cell array of character vectors, then it must be a row of the words of a single document.

Tip

To ensure that the function does not discard useful information, you must first preprocess the input documents using the same steps used to preprocess the documents used to train the model.

Input bag-of-words or bag-of-n-grams model, specified as a bagOfWords object or a bagOfNgrams object. If bag is a bagOfNgrams object, then the function treats each n-gram as a single word.

Frequency counts of words, specified as a matrix of nonnegative integers. If you specify 'DocumentsIn' to be 'rows', then the value counts(i,j) corresponds to the number of times the jth word of the vocabulary appears in the ith document. Otherwise, the value counts(i,j) corresponds to the number of times the ith word of the vocabulary appears in the jth document.

Name-Value Arguments

Specify optional pairs of arguments as Name1=Value1,...,NameN=ValueN, where Name is the argument name and Value is the corresponding value. Name-value arguments must appear after other arguments, but the order of the pairs does not matter.

Before R2021a, use commas to separate each name and value, and enclose Name in quotes.

Example: 'IterationLimit',200 sets the iteration limit to 200.

Note

These name-value pairs only apply if the input model is an ldaModel object.

Orientation of documents in the word count matrix, specified as the comma-separated pair consisting of 'DocumentsIn' and one of the following:

  • 'rows' – Input is a matrix of word counts with rows corresponding to documents.

  • 'columns' – Input is a transposed matrix of word counts with columns corresponding to documents.

This option only applies if you specify the input documents as a matrix of word counts.

Note

If you orient your word count matrix so that documents correspond to columns and specify 'DocumentsIn','columns', then you might experience a significant reduction in optimization-execution time.

Maximum number of iterations, specified as the comma-separated pair consisting of 'IterationLimit' and a positive integer.

Example: 'IterationLimit',200

Relative tolerance on log-likelihood, specified as the comma-separated pair consisting of 'LogLikelihoodTolerance' and a positive scalar. The optimization terminates when this tolerance is reached.

Example: 'LogLikelihoodTolerance',0.001

Output Arguments

collapse all

Output document scores, returned as a matrix of score vectors.

Version History

Introduced in R2017b