0 views

Uploaded by vkv.foe

Machine learning lecture

- s
- A TWO-STAGE HYBRID MODEL BY USING ARTIFICIAL NEURAL NETWORKS AS FEATURE CONSTRUCTION ALGORITHMS
- AI for everyone notes.docx
- [IJCST-V4I6P19]:Shobana.K, Sasikala.M
- Face recognition
- Full and Partial Class Relevant Genes
- Complexity Npc
- OK---Local PCA Regression for Missing Data Estimation in Telecommunication Dataset
- Pca
- 3038847-dzone-2016guidetobigdata
- The effect of Internet usage on Interpersonal Relationships
- MAr
- An Introduction to Swarm Robotics - A.Martinoli_tutorial_slides
- JPMA_998
- Final Year Students Projects Vb Net Java J2ME J2EE Vb Ieee Projects Zebros India
- A Mixture Model of Hubness and Pca
- Data Warehousing Cleaning Phase
- Gayathri Varadarajan Final Thesis Report_final
- NationalStrategy for AI Discussion Paper
- gge

You are on page 1of 28

of our machine learning algorithm.

Selection of data pool is performed in epochs, whereby at each

phase, a batch of one or more queries are performed simultaneously.

batches (and overall data set) from such a pool of candidates makes

direct optimization an NP-hard problem.

promising data greedily one-at a-time from the pool.

what selection criterion is being used), such sequential data access

often works well in practice while making the selection problem

tractable.

NP-hard problems are problems for which there is no known

polynomial algorithm, so that the time to find a solution

grows exponentially with problem size. Although it has not

been definitively proven that, there is no polynomial

algorithm for solving NP-hard problems, many eminent

mathematicians have tried and failed.

Feature construction

Preprocessing.

•

Processing/Feature

construction Steps

Preprocessing

transformation steps

Standardization

comparable objects.

measured in meters and x2 is a height measured in centimeters.

Both can be compared, added or subtracted but it would be

unreasonable to do it before appropriate normalization.

(xi−µi)/σi , where µi and σi are the mean and the standard

deviation of feature xi over training examples.

Normalization

xi ’s are the number of pixels with color i, it makes sense to

normalize x by dividing it by the total number of counts in

order to encode the distribution and remove the

dependence on the size of the image. This translates into

the formula: x’ = x/||x||.

Signal enhancement.

signal or image-processing filters. These operations include

baseline or background removal, de-noising, smoothing, or

sharpening. The Fourier transform and wavelet transforms

are popular methods.

Extraction of local features

techniques like convolutional methods using hand-crafted

kernels or syntactic and structural methods are used. These

techniques encode problem specific knowledge into the

features. They are beyond the scope of this book but it is

worth mentioning that they can bring significant

improvement.

Linear and non-linear space embedding methods

techniques might be used to project or embed the data into

a lower dimensional space while retaining as much

information as possible. Classical examples are Principal

Component Analysis (PCA) and Multidimensional Scaling

(MDS).

space might be used as features or simply as a means of

data visualization.

Non-linear expansions

speaking about complex data, it is sometimes better to

increase the dimensionality. This happens when the problem

is very complex and first order interactions are not enough

to derive good results. This consists for instance in

computing products of the original features xi to create

monomials xk1 xk2 ...xkp .

Non-linear expansions

speaking about complex data, it is sometimes better to

increase the dimensionality. This happens when the problem

is very complex and first order interactions are not enough

to derive good results. This consists for instance in

computing products of the original features xi to create

monomials x_{k1}, x_{k2}, …, x_{kp}.

Feature discretization

makes sense then to discretize continuous values into a

finite discrete set. This step not only facilitates the use of

certain algorithms, it may simplify the data description and

improve data understanding (Liu and Motoda, 1998).

• In particular, one should beware of not losing information

at the feature construction stage.

disadvantage is: : it increases the dimensionality of the

patterns and thereby immerses the relevant information

into a sea of possibly irrelevant, noisy or redundant

features

Feature selection

• It is the process of selecting correct or most informative

features.

• performance improvement

• data understanding

• Feature selection can be performed in following ways:

• Redundant features

Individual relevance ranking

is relevant individually and the other (x2) does not help

providing a better class separation.

feature that provides a good class separation by itself will rank

high and will therefore be chosen.

Rotations in feature space often simplify feature selection,

refer image.

for predicting the class conditional probabilities.

Relevant features that are

individually irrelevant

image that is randomly oﬀset by a local background

change; feature x2 might be measuring such local oﬀset,

which by itself is not informative. Hence, feature x2 might

be completely uncorrelated to the target and yet improve

the separability of feature x1, if subtracted from it.

•

• Two individually irrelevant features may become relevant

when used in combination.

nearest-neighbor algorithm.

between the examples and their nearest misses is

compared to the sum of distances to their nearest hits.

The Relief method works for multi-class problems.

Redundant features

• Noise reduction: When two features provide identical

projected distribution.

dimensional distribution shows a better class separation

than the one achievable with either feature. (refer (d) of

previous figure)

distribution is same but they are not redundant.

- sUploaded byWilliams Danilo Clemente Huanquis
- A TWO-STAGE HYBRID MODEL BY USING ARTIFICIAL NEURAL NETWORKS AS FEATURE CONSTRUCTION ALGORITHMSUploaded byLewis Torres
- AI for everyone notes.docxUploaded byOmer Qureshi
- [IJCST-V4I6P19]:Shobana.K, Sasikala.MUploaded byEighthSenseGroup
- Face recognitionUploaded byChowkidar Karthik
- Full and Partial Class Relevant GenesUploaded bychithrasreemod
- Complexity NpcUploaded byምንሼ ነው ዘመኑ
- OK---Local PCA Regression for Missing Data Estimation in Telecommunication DatasetUploaded byAnonymous RrGVQj
- PcaUploaded byAnil Dongardiye
- 3038847-dzone-2016guidetobigdataUploaded byraulmanzano
- The effect of Internet usage on Interpersonal RelationshipsUploaded byGianina Luția
- MArUploaded bySceptic Granny
- An Introduction to Swarm Robotics - A.Martinoli_tutorial_slidesUploaded byzmaj101
- JPMA_998Uploaded byEric Quanhui
- Final Year Students Projects Vb Net Java J2ME J2EE Vb Ieee Projects Zebros IndiaUploaded byISHAN CHAUDHARY
- A Mixture Model of Hubness and PcaUploaded byAlejandro Carver
- Data Warehousing Cleaning PhaseUploaded byTripathi Vina
- Gayathri Varadarajan Final Thesis Report_finalUploaded bygaya3varad
- NationalStrategy for AI Discussion PaperUploaded byswathy
- ggeUploaded byIsmael Neu
- IJREATV1I2003.pdfUploaded byanon_303010132
- 13 Noor Azman AliUploaded byAlemnkeng Tawung
- Fourth International Conference on Artificial Intelligence, Soft Computing (AISC-2016)Uploaded byCS & IT
- Face Recognition System Hidden Markov ModelUploaded bynab05
- An Intrusion Detection Based on Data Mining Technique and Its IntendedUploaded byIJMTER
- Face Recognition and DetectionUploaded bychaithra580
- Cheminformatics-WekaUploaded byenirahcool
- SivaratMalapet.pdfUploaded bypr
- Cryptocurrency PredictionUploaded byChelmus Rares
- Thornhill&HorchPlantWideDisturbanceCEP2007Uploaded byafdanny

- Bajrang BaanUploaded byvkv.foe
- ML1Uploaded byvkv.foe
- ReadMeUploaded bysarveshweb
- ReadMeUploaded bysarveshweb
- asjkfUploaded byvkv.foe
- Doc1Uploaded byvkv.foe
- Lecture 06Uploaded byvkv.foe
- SampleUploaded byvkv.foe
- SampleUploaded byvkv.foe
- Dfh - NotepadUploaded byvkv.foe
- Topic0 IntroductionUploaded byLakshmideepak Avvaru
- Zafarnama by Guru Gobind Singh Ji (Gurmukhi,Persian,English Meanings)Uploaded bynss1234567890
- Micro ControllersUploaded byvkv.foe

- Export Proengineer Wildfire 4 0 Models to 3d PDFUploaded byTim
- mn00224e-v11-IDU-ALP2-ALCP2-ALCP2e.pdfUploaded byBuzduga Ionut
- Mechanical Engineering Autumn Winter 2013Uploaded bymalmoosawi_1
- desggtUploaded byDevasish Sahoo
- Accenture Digital Transformation in the Age of the CustomerUploaded byNeshvar Dmitri
- test.jsonUploaded byMark Cruz
- DxDiagUploaded byRizki Perdana
- Administrative Coordinator Communications Manager in Raleigh Durham NC Resume Wes SchobelUploaded byWesSchobel
- Cover of Hair DryerUploaded byAlsayed Rabiea Miesalam
- Solid Modeling techniquesUploaded byRa Balamurugan
- A Starlet is Born: New Options for VAX and Alpha ReplacementUploaded byCamiel Vanderhoeven
- NZCS 224 (1).docxUploaded byChristian Chappell
- Quick ScanUploaded byadrianvillalonga
- Class 4 BookletUploaded byZoran Matev
- Enterprise ArchitectureUploaded bythakkarruchita
- 98-368 Test Bank Lesson 05Uploaded byyassinedo
- m2j2setupUploaded byShubham Harnal
- Big-ip Quick Start.pdfUploaded byAdi Dazzy
- MXA6800+_MSA6800+A_Ed-A_175-000347-00Uploaded byTechne Phobos
- Mitutoyo - Twardościomierze Rockwell HR 530 Seria - E117009 - 2017 ENUploaded byD.T.
- kim and kim MALL applications.pdfUploaded byRolf Naidoo
- Survivejs Webpack Apprentice MasterUploaded byCesar Lopez
- DOM2-Core, HTMLUploaded bySundeep Shekhar
- ml505_bsb_std_ip_additionUploaded byNguyễn Hữu Tuyến
- Metrologic MS6220 Pulsar MANUALUploaded byAnita Radukic
- Ez PDF to Word Converter for Doc 5 0 Keygen At4reUploaded byAngelina
- Absence Management Self Service SetupUploaded byParth Desai
- Gyanvriksh4Salesforce.pdfUploaded bydskishore
- User_Manual Intel(R) Matrix Storage ManagerUploaded byNappNew
- Virtual Classroom Tool EditUploaded byRaizzahRapunzelPanis