# All-NBA Predict #22 – Classifying All-NBA Players (Part IV – Logistic Regression)

Okay. So last time, we tried looking at LDA and QDA in addition to our manually chosen linear decision boundary to classify if players are all-NBA team worthy or not. We got the following results: Manually chosen linear decision boundary – 93% / 92% accuracy for all-NBA / non all-NBA respectively LDA – 69% / […]

Read More All-NBA Predict #22 – Classifying All-NBA Players (Part IV – Logistic Regression)

# All-NBA Predict #21 – Classifying All-NBA Players (Part III – Quadratic Discriminant Analysis)

Last post, we reviewed a few different topics Gaussian distribution Multivariate gaussian distribution Linear discriminant analysis We left off feeling a bit unsatisfied with LDA because were were getting 92% / 69% accuracy in prediction for all-NBA players and non all-NBA players respectively. I’m not too fond of that 69%, especially considering that we were […]

Read More All-NBA Predict #21 – Classifying All-NBA Players (Part III – Quadratic Discriminant Analysis)

# All-NBA Predict #20 – Classifying All-NBA Players (Part II – Linear Discriminant Analysis)

Okay, so last time, I selected two boundaries – one a rectangle, one a linear line. Anything on one side of the shape was considered all-NBA, anything on the other was considered non all-NBA. Did it work? For the rectangle, no. Well… I guess I didn’t really try to optimize it, but it was clear […]

Read More All-NBA Predict #20 – Classifying All-NBA Players (Part II – Linear Discriminant Analysis)

# All-NBA Predict #19 – Classifying All-NBA Players (Part I – Manually Selected Boundaries)

Okay, I’ve taken a look at Win Shares and VORP so far, and it looked like there was a pretty clear line between all-nba level players and non all-nba level players. Okay, not a clear line, but there is a… quadrant? Let’s refresh our memories of what the WS vs VORP plot looks like for […]

Read More All-NBA Predict #19 – Classifying All-NBA Players (Part I – Manually Selected Boundaries)

# All-NBA Predict #18 – Exploration of Historical NBA Players (Part VIII, WS & VORP)

Let’s get right into it. Last post, I looked at, well, really I learned what PER, BPM, VORP, and WS were. Basketball-reference’s blurbs now make some sense to me haha: PER – Sum up all a player’s positive accomplishments, subtract the negative accomplishments, and return a per-minute rating of a player’s performance BPM – Box […]

Read More All-NBA Predict #18 – Exploration of Historical NBA Players (Part VIII, WS & VORP)

# All-NBA Predict #17 – Exploration of Historical NBA Players (Part VII, Advanced Metrics)

In post #13, I glossed over advanced metrics and tried to plot them on a PCA plot. Didn’t really work as I didn’t really think it through and I just threw everything up. At one point, I was plotting things like AST% to something like VORP. While we can throw anything on a PCA in […]

Read More All-NBA Predict #17 – Exploration of Historical NBA Players (Part VII, Advanced Metrics)

# All-NBA Predict #16 – Exploration of Historical NBA Players (Part VII, PCA on 2016 Teams)

I’m starting to get a foothold oh leveraging PCA to understand the makeup of a team from their players’ PCA bi-plots. It’s really been a useful tool to see where teams were strong, where teams were weak, and the overall balance of a team. Like I left off last time, I’d now like to look […]

Read More All-NBA Predict #16 – Exploration of Historical NBA Players (Part VII, PCA on 2016 Teams)

# All-NBA Predict #15 – Exploration of Historical NBA Players (Part VI, PCA to Explore Historical Lineups)

Last time, I explored 9 regions of our PCA bi-plot. I’ll review what I thought the 9 regions represented here: Region 1: All-Star Guards Region 2: All-Star Scorers Region 3: All-Star Big Men Region 4: Great Guards Region 5: Great All-Around / Two-Way Players Region 6: Great Big Men Region 7: Spot Up Shooters Region […]

Read More All-NBA Predict #15 – Exploration of Historical NBA Players (Part VI, PCA to Explore Historical Lineups)

# All-NBA Predict #14 – Exploration of Historical NBA Players (Part V, PCA to Cluster Playing Styles)

In our last post, we harnessed the power of PCA even further to look at different ways to break up the PCA bi-plot. With ggbiplot’s “ellipse” argument, we were able to check out where different players lie on the plot. To review, my 3 main questions were: How closely can I map different “types” of […]

Read More All-NBA Predict #14 – Exploration of Historical NBA Players (Part V, PCA to Cluster Playing Styles)

# All-NBA Predict #13 – Exploration of Historical NBA Players (Part IV, PCA on Advanced Metrics & By Era)

Advanced Metrics Okay, so at this point, I’ve seen some of the benefits of data exploration methods. I really liked the PCA bi-plot for reasons I rambled about in the last post, namely the balance between interpretability and the value of information provided. Here, I’m just going to throw it at some of the advanced […]

Read More All-NBA Predict #13 – Exploration of Historical NBA Players (Part IV, PCA on Advanced Metrics & By Era)