search

bensivo / mapo

0 stars 0 forks source link

Bug: Editing text node causes node to shift #99

Open bensivo opened 3 weeks ago

bensivo commented 3 weeks ago

Image

.mapo file

{"id":null,"name":"CS441 2.1 Classification","edges":[{"id":"b4f3d8d8-8f06-47d0-83a4-eafaef011d03","startNodeId":"7f6efb93-2075-4ae6-b4d2-3f4bca8a9b15","endNodeId":"34033f22-8f24-4fbb-90bc-4b7ba0b75f07","text":"use"},{"id":"acac5b48-34ac-4ec4-85ce-c08e1633154a","startNodeId":"51b212de-b579-4bad-914a-7cba738bac39","endNodeId":"7f6efb93-2075-4ae6-b4d2-3f4bca8a9b15","text":"is a"},{"id":"c5b226ae-52f2-47a3-bdb5-dab52f06ff4e","startNodeId":"64b9020c-8281-4a31-b39f-2ddfdc1cd417","endNodeId":"7f6efb93-2075-4ae6-b4d2-3f4bca8a9b15","text":"is a"},{"id":"cfa09a61-0a00-49aa-a187-05f993870eef","startNodeId":"3450e605-1b1f-459a-8e64-33f085ad6727","endNodeId":"34033f22-8f24-4fbb-90bc-4b7ba0b75f07","text":"Created from"},{"id":"ddc5dec1-147d-46c9-93d6-9bb515a787de","startNodeId":"cab071e2-d7ef-4250-9eb4-3c3bb942fb34","endNodeId":"b6bd042b-d3ef-49a8-a90f-e7232bddbb9a","text":"kind of"},{"id":"47cab9bc-cd16-4985-8018-282e734d8388","startNodeId":"112029e3-0859-4ec4-8e84-66b8e6f1447f","endNodeId":"b6bd042b-d3ef-49a8-a90f-e7232bddbb9a","text":"kind of"},{"id":"f0175439-a5b9-469e-b51f-54f981e7ea78","startNodeId":"b6bd042b-d3ef-49a8-a90f-e7232bddbb9a","endNodeId":"95cd61f5-0986-4fb8-a0b8-22492d4f2145","text":"mitigates"},{"id":"7140b1a5-d21a-4935-a496-013b00276fdd","startNodeId":"0da4b816-d1bb-4ca2-94d4-807cc2ff26c5","endNodeId":"1d9c04c4-5aaf-40a1-b327-6c521f1d2720","text":"extends"}],"textNodes":[{"id":"34033f22-8f24-4fbb-90bc-4b7ba0b75f07","text":"Evaluating classifiers (2.1)\n\nTP = Said 1 and it was 1\nFP = Said 1, but it was 0\nFN = Said 0 but it was 1\nTN = Said 0 and it was 0","x":582.5,"y":462.46000000000004},{"id":"7f6efb93-2075-4ae6-b4d2-3f4bca8a9b15","text":"Performance Metrics (2.1)\n\nAccuracy: number correct guesses / total guesses\nError Rate: incorrect guesses / total guesses","x":272.5,"y":746.46},{"id":"51b212de-b579-4bad-914a-7cba738bac39","text":"Sensitivity\nAccuracy for all the positive guesses\n\nTP/(TP+FN)","x":97.5,"y":995.46},{"id":"64b9020c-8281-4a31-b39f-2ddfdc1cd417","text":"Specificity\nAccuracy for all the negative guesses\n\nTN/(TN+FP)","x":451.5,"y":994.46},{"id":"3450e605-1b1f-459a-8e64-33f085ad6727","text":"Class Confusion Matrix\n\nRows = True labels\nColumns = Predicted labels\nValue = no. of items in that category ","x":795.5,"y":741.46},{"id":"b6bd042b-d3ef-49a8-a90f-e7232bddbb9a","text":"Cross-Validation (2.1)\n\nGenerate a new \"fold\" of training and test set on each iter\nAverage accuracy for all folds","x":1265.735738410613,"y":417.2175233482208},{"id":"cab071e2-d7ef-4250-9eb4-3c3bb942fb34","text":"Leave-One-Out Cross-Validation (2.1)\n\nSpecial case of Cross Validation\nTraining set in each fold = entire set except 1 item\nTest set = 1 item","x":1249.5,"y":676.46},{"id":"112029e3-0859-4ec4-8e84-66b8e6f1447f","text":"K-Fold Cross-Validation (2.1)\n\nDataset is split into k pieces, at fixed places\nTraining set at each iter is one of these k pieces\n\nIter 1:    TEST | N/K | N/K | ...\nIter 2:     N/K | TEST | N/K | ...","x":1782.8911377322945,"y":691.9793798013914},{"id":"95cd61f5-0986-4fb8-a0b8-22492d4f2145","text":"Overfitting","x":1519.9333759771418,"y":208.4763985096585},{"id":"1d9c04c4-5aaf-40a1-b327-6c521f1d2720","text":"Nearest Neighbor\n\n\"Training\": save vectors for all items in training set\nClassification: calc dist to all saved vectors, assign label from nearest\n\nIssue: classiification can change for every single change to training data","x":239.62151647748408,"y":-264.8402817751638},{"id":"0da4b816-d1bb-4ca2-94d4-807cc2ff26c5","text":"K-Nearest Neighbor\n\nFind K closest items to input vector, return most common label\nSlightly more robust","x":-185.37698846549,"y":9.103914721146168}]}
bensivo commented 3 weeks ago

Seems to only consistently affect the k-nearest neighbor node. Not the others?