feat u re e x traction
play

Feat u re e x traction D IME N SION AL ITY R E D U C TION IN P YTH - PowerPoint PPT Presentation

Sep 02, 2022 •359 likes •946 views

Feat u re e x traction D IME N SION AL ITY R E D U C TION IN P YTH ON Jeroen Boe y e Machine Learning Engineer , Faktion Feat u re selection DIMENSIONALITY REDUCTION IN PYTHON Feat u re selection Feat u re e x traction DIMENSIONALITY REDUCTION

  1. Feat u re e x traction D IME N SION AL ITY R E D U C TION IN P YTH ON Jeroen Boe y e Machine Learning Engineer , Faktion

  2. Feat u re selection DIMENSIONALITY REDUCTION IN PYTHON

  3. Feat u re selection Feat u re e x traction DIMENSIONALITY REDUCTION IN PYTHON

  4. Feat u re generation - BMI df_body['BMI'] = df_body['Weight kg'] / df_body['Height m'] ** 2 DIMENSIONALITY REDUCTION IN PYTHON

  5. Feat u re generation - BMI df_body['BMI'] = df_body['Weight kg'] / df_body['Height m'] ** 2 Weight kg Height m BMI 81.5 1.776 25.84 72.6 1.702 25.06 92.9 1.735 30.86 DIMENSIONALITY REDUCTION IN PYTHON

  6. Feat u re generation - BMI df_body.drop(['Weight kg', 'Height m'], axis=1) BMI 25.84 25.06 30.86 DIMENSIONALITY REDUCTION IN PYTHON

  7. Feat u re generation - a v erages le � leg mm right leg mm 882 885 870 869 901 900 leg_df['leg mm'] = leg_df[['right leg mm', 'left leg mm']].mean(axis=1) DIMENSIONALITY REDUCTION IN PYTHON

  8. Feat u re generation - a v erages leg_df.drop(['right leg mm', 'left leg mm'], axis=1) leg mm 883.5 869.5 900.5 DIMENSIONALITY REDUCTION IN PYTHON

  9. Cost of taking the a v erage DIMENSIONALITY REDUCTION IN PYTHON

  10. Cost of taking the a v erage DIMENSIONALITY REDUCTION IN PYTHON

  11. Cost of taking the a v erage DIMENSIONALITY REDUCTION IN PYTHON

  12. Cost of taking the a v erage DIMENSIONALITY REDUCTION IN PYTHON

  13. Intro to PCA sns.scatterplot(data=df, x='handlength', y='footlength') DIMENSIONALITY REDUCTION IN PYTHON

  14. Intro to PCA scaler = StandardScaler() df_std = pd.DataFrame(scaler.fit_transform(df), columns = df.columns) DIMENSIONALITY REDUCTION IN PYTHON

  15. Intro to PCA scaler = StandardScaler() df_std = pd.DataFrame(scaler.fit_transform(df), columns = df.columns) DIMENSIONALITY REDUCTION IN PYTHON

  16. Intro to PCA scaler = StandardScaler() df_std = pd.DataFrame(scaler.fit_transform(df), columns = df.columns) DIMENSIONALITY REDUCTION IN PYTHON

  17. Intro to PCA scaler = StandardScaler() df_std = pd.DataFrame(scaler.fit_transform(df), columns = df.columns) DIMENSIONALITY REDUCTION IN PYTHON

  18. Let ' s practice ! D IME N SION AL ITY R E D U C TION IN P YTH ON

  19. Principal component anal y sis D IME N SION AL ITY R E D U C TION IN P YTH ON Jeroen Boe y e Machine Learning Engineer , Faktion

  20. PCA concept DIMENSIONALITY REDUCTION IN PYTHON

  21. PCA concept DIMENSIONALITY REDUCTION IN PYTHON

  22. PCA concept DIMENSIONALITY REDUCTION IN PYTHON

  23. Calc u lating the principal components from sklearn.preprocessing import StandardScaler scaler = StandardScaler() std_df = scaler.fit_transform(df) from sklearn.decomposition import PCA pca = PCA() print(pca.fit_transform(std_df)) [[-0.08320426 -0.12242952] [ 0.31478004 0.57048158] ... [-0.5609523 0.13713944] [-0.0448304 -0.37898246]] DIMENSIONALITY REDUCTION IN PYTHON

  24. PCA remo v es correlation DIMENSIONALITY REDUCTION IN PYTHON

  25. Principal component e x plained v ariance ratio from sklearn.decomposition import PCA pca = PCA() pca.fit(std_df) print(pca.explained_variance_ratio_) array([0.90, 0.10]) DIMENSIONALITY REDUCTION IN PYTHON

  26. PCA for dimensionalit y red u ction DIMENSIONALITY REDUCTION IN PYTHON

  27. PCA for dimensionalit y red u ction print(pca.explained_variance_ratio_) array([0.9997, 0.0003]) DIMENSIONALITY REDUCTION IN PYTHON

  28. PCA for dimensionalit y red u ction pca = PCA() pca.fit(ansur_std_df) print(pca.explained_variance_ratio_) array([0.44, 0.18, 0.04, 0.03, 0.02, 0.02, 0.02, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , ... 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. , 0. ]) DIMENSIONALITY REDUCTION IN PYTHON

  29. PCA for dimensionalit y red u ction pca = PCA() pca.fit(ansur_std_df) print(pca.explained_variance_ratio_.cumsum()) array([0.44, 0.62, 0.66, 0.69, 0.72, 0.74, 0.76, 0.77, 0.79, 0.8 , 0.81, 0.82, 0.83, 0.84, 0.85, 0.86, 0.87, 0.87, 0.88, 0.89, 0.89, 0.9 , 0.9 , 0.91, 0.92, 0.92, 0.92, 0.93, 0.93, 0.94, 0.94, 0.94, 0.95, ... 0.99, 0.99, 0.99, 0.99, 0.99, 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. ]) DIMENSIONALITY REDUCTION IN PYTHON

  30. Let ' s practice ! D IME N SION AL ITY R E D U C TION IN P YTH ON

  31. PCA applications D IME N SION AL ITY R E D U C TION IN P YTH ON Jeroen Boe y e Machine Learning Engineer , Faktion

  32. Understanding the components print(pca.components_) array([[ 0.71, 0.71], [ -0.71, 0.71]]) PC 1 = 0.71 x Hand length + 0.71 x Foot length PC 2 = -0.71 x Hand length + 0.71 x Foot length DIMENSIONALITY REDUCTION IN PYTHON

  33. PCA for data e x ploration DIMENSIONALITY REDUCTION IN PYTHON

  34. PCA in a pipeline from sklearn.preprocessing import StandardScaler from sklearn.decomposition import PCA from sklearn.pipeline import Pipeline pipe = Pipeline([ ('scaler', StandardScaler()), ('reducer', PCA())]) pc = pipe.fit_transform(ansur_df) print(pc[:,:2]) array([[-3.46114925, 1.5785215 ], [ 0.90860615, 2.02379935], ..., [10.7569818 , -1.40222755], [ 7.64802025, 1.07406209]]) DIMENSIONALITY REDUCTION IN PYTHON

  35. Checking the effect of categorical feat u res print(ansur_categories.head()) Branch Component Gender BMI _ class Height _ class Combat Arms Reg u lar Arm y Male O v er w eight Tall Combat S u pport Reg u lar Arm y Male O v er w eight Normal Combat S u pport Reg u lar Arm y Male O v er w eight Normal Combat Ser v ice S u pport Reg u lar Arm y Male O v er w eight Normal Combat Ser v ice S u pport Reg u lar Arm y Male O v er w eight Tall DIMENSIONALITY REDUCTION IN PYTHON

  36. Checking the effect of categorical feat u res ansur_categories['PC 1'] = pc[:,0] ansur_categories['PC 2'] = pc[:,1] sns.scatterplot(data=ansur_categories, x='PC 1', y='PC 2', hue='Height_class', alpha=0.4) DIMENSIONALITY REDUCTION IN PYTHON

  37. Checking the effect of categorical feat u res sns.scatterplot(data=ansur_categories, x='PC 1', y='PC 2', hue='Gender', alpha=0.4) DIMENSIONALITY REDUCTION IN PYTHON

  38. Checking the effect of categorical feat u res sns.scatterplot(data=ansur_categories, x='PC 1', y='PC 2', hue='BMI_class', alpha=0.4 DIMENSIONALITY REDUCTION IN PYTHON

  39. PCA in a model pipeline pipe = Pipeline([ ('scaler', StandardScaler()), ('reducer', PCA(n_components=3)), ('classifier', RandomForestClassifier())]) pipe.fit(X_train, y_train) print(pipe.steps[1]) ('reducer', PCA(copy=True, iterated_power='auto', n_components=3, random_state=None, svd_solver='auto', tol=0.0, whiten=False)) DIMENSIONALITY REDUCTION IN PYTHON

  40. PCA in a model pipeline pipe.steps[1][1].explained_variance_ratio_.cumsum() array([0.56, 0.69, 0.74]) print(pipe.score(X_test, y_test)) 0.986 DIMENSIONALITY REDUCTION IN PYTHON

  41. Let ' s practice ! D IME N SION AL ITY R E D U C TION IN P YTH ON

  42. Principal Component selection D IME N SION AL ITY R E D U C TION IN P YTH ON Jeroen Boe y e Machine Learning Engineer , Faktion

  43. Setting an e x plained v ariance threshold pipe = Pipeline([ ('scaler', StandardScaler()), ('reducer', PCA(n_components=0.9))]) # Fit the pipe to the data pipe.fit(poke_df) print(len(pipe.steps[1][1].components_)) 5 DIMENSIONALITY REDUCTION IN PYTHON

  44. An optimal n u mber of components pipe.fit(poke_df) var = pipe.steps[1][1].explained_variance_ratio_ plt.plot(var) plt.xlabel('Principal component index') plt.ylabel('Explained variance ratio') plt.show() DIMENSIONALITY REDUCTION IN PYTHON

  45. An optimal n u mber of components pipe.fit(poke_df) var = pipe.steps[1][1].explained_variance_ratio_ plt.plot(var) plt.xlabel('Principal component index') plt.ylabel('Explained variance ratio') plt.show() DIMENSIONALITY REDUCTION IN PYTHON

  46. PCA operations DIMENSIONALITY REDUCTION IN PYTHON

  47. PCA operations DIMENSIONALITY REDUCTION IN PYTHON

  48. PCA operations DIMENSIONALITY REDUCTION IN PYTHON

  49. Compressing images DIMENSIONALITY REDUCTION IN PYTHON

  50. Compressing images print(X_test.shape) (15, 2914) 62 x 47 pi x els = 2914 gra y scale v al u es print(X_train.shape) (1333, 2914) DIMENSIONALITY REDUCTION IN PYTHON

  51. Compressing images pipe = Pipeline([ ('scaler', StandardScaler()), ('reducer', PCA(n_components=290))]) pipe.fit(X_train) pc = pipe.fit_transform(X_test) print(pc.shape) (15, 290) DIMENSIONALITY REDUCTION IN PYTHON

  52. Reb u ilding images pc = pipe.transform(X_test) print(pc.shape) (15, 290) X_rebuilt = pipe.inverse_transform(pc) print(X_rebuilt.shape) (15, 2914) img_plotter(X_rebuilt) DIMENSIONALITY REDUCTION IN PYTHON

  53. Reb u ilding images DIMENSIONALITY REDUCTION IN PYTHON

  54. Let ' s practice ! D IME N SION AL ITY R E D U C TION IN P YTH ON

  55. Congrat u lations ! D IME N SION AL ITY R E D U C TION IN P YTH ON Jeroen Machine Learning Engineer , Faktion

Recommend

Dimensionalit y red u ction : feat u re e x traction P R AC TIC IN G MAC H IN E L E AR N IN G

Dimensionalit y red u ction : feat u re e x traction P R AC TIC IN G MAC H IN E L E AR N IN G

Dimensionalit y red u ction : feat u re e x traction P R AC TIC IN G MAC H IN E L E AR N IN G IN TE R VIE W QU E STION S IN P YTH ON Lisa St u art Data Scientist Uns u per v ised learning methods Principal component anal y sis ( PCA ) -->

426 views • 41 slides
Feat u re engineering P R E P R OC E SSIN G FOR MAC H IN E L E AR N IN G IN P YTH ON Sarah G

Feat u re engineering P R E P R OC E SSIN G FOR MAC H IN E L E AR N IN G IN P YTH ON Sarah G

Feat u re engineering P R E P R OC E SSIN G FOR MAC H IN E L E AR N IN G IN P YTH ON Sarah G u ido Senior Data Scientist What is feat u re engineering ? Creation of ne w feat u res based on e x isting feat u res Insight into relationships bet

680 views • 23 slides
Feat ure Select ion using/ f or Feat ure Select ion using/ f or Transduct ive ransduct ive S

Feat ure Select ion using/ f or Feat ure Select ion using/ f or Transduct ive ransduct ive S

NI PS03 Workshop of Feature Extraction and Feature Selection Challenge Feat ure Select ion using/ f or Feat ure Select ion using/ f or Transduct ive ransduct ive S Support upport V Vect or ect or M Machine achine T Mr. Zhili Wu Dr.

294 views • 13 slides
Wh y generate feat u res ? FE ATU R E E N G IN E E R IN G FOR MAC H IN E L E AR N IN G IN P

Wh y generate feat u res ? FE ATU R E E N G IN E E R IN G FOR MAC H IN E L E AR N IN G IN P

Wh y generate feat u res ? FE ATU R E E N G IN E E R IN G FOR MAC H IN E L E AR N IN G IN P YTH ON Robert O ' Callaghan Director of Data Science , Ordergroo v e Feat u re Engineering FEATURE ENGINEERING FOR MACHINE LEARNING IN PYTHON

461 views • 29 slides
Regression : feat u re selection P R AC TIC IN G MAC H IN E L E AR N IN G IN TE R VIE W QU E

Regression : feat u re selection P R AC TIC IN G MAC H IN E L E AR N IN G IN TE R VIE W QU E

Regression : feat u re selection P R AC TIC IN G MAC H IN E L E AR N IN G IN TE R VIE W QU E STION S IN P YTH ON Lisa St u art Data Scientist Selecting the correct feat u res : Red u ces o v er ing Impro v es acc u rac y Increases

859 views • 54 slides
Feat u re Generation FE ATU R E E N G IN E E R IN G W ITH P YSPAR K John Hog u e Lead Data

Feat u re Generation FE ATU R E E N G IN E E R IN G W ITH P YSPAR K John Hog u e Lead Data

Feat u re Generation FE ATU R E E N G IN E E R IN G W ITH P YSPAR K John Hog u e Lead Data Scientist Wh y generate ne w feat u res ? M u ltipl y ing S u mming Di erencing Di v iding FEATURE ENGINEERING WITH PYSPARK Wh y generate ne w feat

704 views • 34 slides
Traction Getting Traction for (your) Open Source Projects Michael Boelen

Traction Getting Traction for (your) Open Source Projects Michael Boelen

Traction Getting Traction for (your) Open Source Projects Michael Boelen michael.boelen@cisofy.com T-DOSE 2016, 12 November (NLLGG track) Why? (developers) Promote your open source project Users Feedback Invisible benefits 2

674 views • 49 slides
Feat u re crossing FE ATU R E E N G IN E E R IN G IN R Jose Hernande z Data Scientist , Uni v

Feat u re crossing FE ATU R E E N G IN E E R IN G IN R Jose Hernande z Data Scientist , Uni v

Feat u re crossing FE ATU R E E N G IN E E R IN G IN R Jose Hernande z Data Scientist , Uni v ersit y of Washington E x amples of crossing feat u res Ho u sing price prediction [ Location x bedroom number ] H u man a rib u tes [ gender x

958 views • 34 slides
HIGH POWER IGBT TRACTION DRIVES Marc DEBRUYNE Master Expert Traction Systems NIX Converter

HIGH POWER IGBT TRACTION DRIVES Marc DEBRUYNE Master Expert Traction Systems NIX Converter

November 2001 HIGH POWER IGBT TRACTION DRIVES Marc DEBRUYNE Master Expert Traction Systems NIX Converter range 4000 ONIX 3000 Converter Voltage (V) 2000 ONIX 1500 1000 ONIX 800 500 Onix 350 100 200 500 1000 1500 Converter Power

399 views • 21 slides
- African Traction Systems SAIMENA - The South African Institute of Marine Engineers & Naval

- African Traction Systems SAIMENA - The South African Institute of Marine Engineers & Naval

Diesel Electric Propulsion Systems and Controls - African Traction Systems SAIMENA - The South African Institute of Marine Engineers & Naval Architects Presented by Wicus du Preez African Traction Systems A frican T raction S ystems 2015

553 views • 39 slides
E-34028 Replacement of Traction Elevators Clark Street Station September 16, 2019 Project

E-34028 Replacement of Traction Elevators Clark Street Station September 16, 2019 Project

E-34028 Replacement of Traction Elevators Clark Street Station September 16, 2019 Project Scope Major Items Replace three traction elevators Replace all equipment in multi-level elevator machine rooms Provide new controllers,

339 views • 9 slides
MACCON Parameter studies for the optimization of e-mobility traction motors with the help of

MACCON Parameter studies for the optimization of e-mobility traction motors with the help of

MACCON Parameter studies for the optimization of e-mobility traction motors with the help of FluxMotor, FLUX and HyperStudy Dipl.-Ing. Christoph Stuckmann 1 Parameter studies for the optimization of e-mobility traction motors , Altair 2. EM

598 views • 24 slides
TRACTION POWER DESIGN TP Simulation Study- Simulation runs for South and North Corridor.

TRACTION POWER DESIGN TP Simulation Study- Simulation runs for South and North Corridor.

TRACTION POWER DESIGN TP Simulation Study- Simulation runs for South and North Corridor. Prefabricated Traction Power Substations furnished by SANDAG: 13 NEW SUBSTATIONS Existing Substations(Modifications by Mid-Coast Contract)

710 views • 22 slides
Selecting feat u res for model performance D IME N SION AL ITY R E D U C TION IN P YTH ON

Selecting feat u res for model performance D IME N SION AL ITY R E D U C TION IN P YTH ON

Selecting feat u res for model performance D IME N SION AL ITY R E D U C TION IN P YTH ON Jeroen Boe y e Machine Learning Engineer , Faktion Ans u r dataset sample DIMENSIONALITY REDUCTION IN PYTHON Pre - processing the data from

710 views • 44 slides
Transforming ne w feat u res FE ATU R E E N G IN E E R IN G IN R Jose Hernande z Data Scientist

Transforming ne w feat u res FE ATU R E E N G IN E E R IN G IN R Jose Hernande z Data Scientist

Transforming ne w feat u res FE ATU R E E N G IN E E R IN G IN R Jose Hernande z Data Scientist , Uni v ersit y of Washington Addressing ske w ed v ariables ggplot(online_retail, aes(x = Quantity)) + geom_density() FEATURE ENGINEERING IN R Po

475 views • 29 slides
Classification and feat u re engineering MAC H IN E L E AR N IN G FOR TIME SE R IE S DATA IN

Classification and feat u re engineering MAC H IN E L E AR N IN G FOR TIME SE R IE S DATA IN

Classification and feat u re engineering MAC H IN E L E AR N IN G FOR TIME SE R IE S DATA IN P YTH ON Chris Holdgraf Fello w, Berkele y Instit u te for Data Science Al w a y s v is u ali z e ra w data before fitting models MACHINE LEARNING

579 views • 37 slides
Writing Well 2 : Building Traction and Triumph into co-Authorship Suzanne M. Kresta, John A.

Writing Well 2 : Building Traction and Triumph into co-Authorship Suzanne M. Kresta, John A.

Writing Well 2 : Building Traction and Triumph into co-Authorship Suzanne M. Kresta, John A. Nychka Roger Graves*, Jacob Masliyah, Murray Gray * Director, Writing Across the Curriculum, English and Film Studies Chemic Chemical and al and

242 views • 20 slides
MRT9 Specifications Features Feat es Performance Performance Engine Power with

MRT9 Specifications Features Feat es Performance Performance Engine Power with

MRT9 Specifications Features Feat es Performance Performance Engine Power with 19mm restrictor Yamaha YZF-600R, 4 cylinder, 600cc 100 bhp at the flywheel Garrett GT-12 Turbocharger E-85 Bio-ethanol fuel

308 views • 4 slides
Severs Disease Traction apophysitis at the Achilles Tendon Insertion Presentation 12 year

Severs Disease Traction apophysitis at the Achilles Tendon Insertion Presentation 12 year

Severs Disease Traction apophysitis at the Achilles Tendon Insertion Presentation 12 year old male Rugby player Extremely hard playing surfaces due to drought Pain - posterio-plantar aspect of calcaneus Negative Lunge test

745 views • 11 slides
traction options for Far North lines David Shirres, Editor Responding to the environmental

traction options for Far North lines David Shirres, Editor Responding to the environmental

Zero-carbon rail traction options for Far North lines David Shirres, Editor Responding to the environmental agenda Public concern about issues such as climate change and the impact of business on society has never been more intense than it

654 views • 20 slides
Time - dela y ed feat u res and a u to - regressi v e models MAC H IN E L E AR N IN G FOR TIME

Time - dela y ed feat u res and a u to - regressi v e models MAC H IN E L E AR N IN G FOR TIME

Time - dela y ed feat u res and a u to - regressi v e models MAC H IN E L E AR N IN G FOR TIME SE R IE S DATA IN P YTH ON Chris Holdgraf Fello w, Berkele y Instit u te for Data Science The past is u sef u l Timeseries data almost al w a y s

624 views • 48 slides
Presentation August 2020 Ca Cannabis is 3.0 .0 1.0 .0 Growers 2.0 .0 Extr traction 3.0

Presentation August 2020 Ca Cannabis is 3.0 .0 1.0 .0 Growers 2.0 .0 Extr traction 3.0

Corporate Presentation August 2020 Ca Cannabis is 3.0 .0 1.0 .0 Growers 2.0 .0 Extr traction 3.0 .0 Craft We estimate there are ~300,000 Craft Growers in North America alone. Worldwide it is a huge market horizontal. New

394 views • 16 slides
CH CHECKLIST ECKLIST of f Essential sential Fe Feat atures ures Medium-to-High Densities

CH CHECKLIST ECKLIST of f Essential sential Fe Feat atures ures Medium-to-High Densities

RIAN & TR PE PEDE DEST STRIAN TRAN ANSIT SIT-ORIEN ORIENTED TED DE DESIGN SIGN Reid Ewing and Keith Bartholomew Ch Chapt apter er 3 CH CHECKLIST ECKLIST of f Essential sential Fe Feat atures ures Medium-to-High

618 views • 24 slides
Design Procedure of IPM Motor Drive for Railway Traction Massimo Barcaro Emanuele Fornasiero

Design Procedure of IPM Motor Drive for Railway Traction Massimo Barcaro Emanuele Fornasiero

Design Procedure of IPM Motor Drive for Railway Traction Massimo Barcaro Emanuele Fornasiero Nicola Bianchi Silverio Bolognani Electric Drives Laboratory Department of Electrical Engineering University of Padova IEEE - IEMDC 2011

624 views • 37 slides

More recommend