diff --git a/js/view/gan.js b/js/view/gan.js
index 465ea725..e742e229 100644
--- a/js/view/gan.js
+++ b/js/view/gan.js
@@ -123,7 +123,14 @@ export default function (platform) {
 		{ name: 'dis_rate', title: 'D', value: 0.5 },
 	]) {
 		const grd = ganRatesDiv.div()
-		rateElms[v.name] = grd.input.number({ label: v.title, name: v.name, min: 0, max: 100, step: 0.01, value: v.value })
+		rateElms[v.name] = grd.input.number({
+			label: v.title,
+			name: v.name,
+			min: 0,
+			max: 100,
+			step: 0.01,
+			value: v.value,
+		})
 	}
 	const batch = controller.input.number({ label: ' Batch size ', min: 1, max: 100, value: 10 })
 	let threshold = null
diff --git a/tests/lib/model/categorical_naive_bayes.test.js b/tests/lib/model/categorical_naive_bayes.test.js
index 45bee0d9..4e87da9f 100644
--- a/tests/lib/model/categorical_naive_bayes.test.js
+++ b/tests/lib/model/categorical_naive_bayes.test.js
@@ -32,3 +32,32 @@ test('predict', () => {
 	const acc = accuracy(y, t)
 	expect(acc).toBeGreaterThan(0.95)
 })
+
+test('predict fit twice', () => {
+	const model = new CategoricalNaiveBayes()
+	const x = [['a']]
+	const t = []
+	for (let i = 0; i < x.length; i++) {
+		t[i] = 'x'
+	}
+
+	model.fit(x, t)
+	model.fit(x, t)
+
+	const y = model.predict([['a']])
+	expect(y).toEqual(['x'])
+})
+
+test('predict unknown data label', () => {
+	const model = new CategoricalNaiveBayes()
+	const x = [['a']]
+	const t = []
+	for (let i = 0; i < x.length; i++) {
+		t[i] = 'x'
+	}
+
+	model.fit(x, t)
+
+	const y = model.predict([['b']])
+	expect(y).toEqual([null])
+})
diff --git a/tests/lib/model/catmull_rom.test.js b/tests/lib/model/catmull_rom.test.js
index 3d8b529f..debc2495 100644
--- a/tests/lib/model/catmull_rom.test.js
+++ b/tests/lib/model/catmull_rom.test.js
@@ -27,23 +27,35 @@ test('CatmullRomSplines', () => {
 	expect(err).toBeLessThan(0.1)
 })
 
-test('CentripetalCatmullRomSplines', () => {
-	const model = new CentripetalCatmullRomSplines()
-	const x = Matrix.random(20, 1, -2, 2).value
-	const t = []
-	for (let i = 0; i < x.length; i++) {
-		t[i] = Math.sin(x[i])
-	}
-	model.fit(x, t)
-
-	const y = model.predict(x)
-	expect(y).toHaveLength(x.length)
-	for (let i = 0; i < y.length; i++) {
-		expect(y[i]).toBeCloseTo(t[i])
-	}
-
-	const x0 = Matrix.random(100, 1, -2, 2).value
-	const y0 = model.predict(x0)
-	const err = rmse(y0, x0.map(Math.sin))
-	expect(err).toBeLessThan(0.1)
+describe('CentripetalCatmullRomSplines', () => {
+	test('default', () => {
+		const model = new CentripetalCatmullRomSplines()
+		const x = Matrix.random(20, 1, -2, 2).value
+		const t = []
+		for (let i = 0; i < x.length; i++) {
+			t[i] = Math.sin(x[i])
+		}
+		model.fit(x, t)
+
+		const y = model.predict(x)
+		expect(y).toHaveLength(x.length)
+		for (let i = 0; i < y.length; i++) {
+			expect(y[i]).toBeCloseTo(t[i])
+		}
+
+		const x0 = Matrix.random(100, 1, -2, 2).value
+		const y0 = model.predict(x0)
+		const err = rmse(y0, x0.map(Math.sin))
+		expect(err).toBeLessThan(0.1)
+	})
+
+	test('predict second largest value', () => {
+		const model = new CentripetalCatmullRomSplines()
+		const x = [0, 2]
+		const t = [0, 1]
+		model.fit(x, t)
+
+		const y = model.predict([1])
+		expect(y[0]).toBeCloseTo(0.5)
+	})
 })
diff --git a/tests/lib/model/change_finder.test.js b/tests/lib/model/change_finder.test.js
index b7057c35..04eb239d 100644
--- a/tests/lib/model/change_finder.test.js
+++ b/tests/lib/model/change_finder.test.js
@@ -1,11 +1,22 @@
 import Matrix from '../../../lib/util/matrix.js'
 import ChangeFinder from '../../../lib/model/change_finder.js'
 
-test('anomaly detection', () => {
-	const model = new ChangeFinder(5)
-	const n = 50
-	const x = Matrix.concat(Matrix.random(n, 1, 0, 1), Matrix.random(n, 1, 10, 11)).value
-	model.fit(x)
-	const p = model.predict()
-	expect(p).toHaveLength(100)
+describe('anomaly detection', () => {
+	test('default', () => {
+		const model = new ChangeFinder()
+		const n = 50
+		const x = Matrix.concat(Matrix.random(n, 1, 0, 1), Matrix.random(n, 1, 10, 11)).value
+		model.fit(x)
+		const p = model.predict()
+		expect(p).toHaveLength(100)
+	})
+
+	test('with params', () => {
+		const model = new ChangeFinder(5)
+		const n = 50
+		const x = Matrix.concat(Matrix.random(n, 1, 0, 1), Matrix.random(n, 1, 10, 11)).value
+		model.fit(x)
+		const p = model.predict()
+		expect(p).toHaveLength(100)
+	})
 })
diff --git a/tests/lib/model/co_training.test.js b/tests/lib/model/co_training.test.js
index f1ac23d2..ea219b2f 100644
--- a/tests/lib/model/co_training.test.js
+++ b/tests/lib/model/co_training.test.js
@@ -77,3 +77,33 @@ test('semi-classifier', () => {
 	const acc = accuracy(y, t_org)
 	expect(acc).toBeGreaterThan(0.95)
 })
+
+test('semi-classifier not categorized', () => {
+	const model = new CoTraining(
+		{
+			fit(x, y) {},
+			predict(x) {
+				return x.map(v => ({ category: 0, score: 0 }))
+			},
+			threshold: 0.5,
+		},
+		{
+			fit(x, y) {},
+			predict(x) {
+				return x.map(() => ({ category: 0, score: 0 }))
+			},
+			threshold: 0.5,
+		}
+	)
+	const x = [
+		[0, 0],
+		[1, 1],
+	]
+	const t = [0, null]
+	model.init(x, t)
+	for (let i = 0; i < 1; i++) {
+		model.fit()
+	}
+	const y = model.predict(x)
+	expect(y).toEqual([0, null])
+})
diff --git a/tests/lib/model/confidence_weighted.test.js b/tests/lib/model/confidence_weighted.test.js
index 61e445ff..b28dcc74 100644
--- a/tests/lib/model/confidence_weighted.test.js
+++ b/tests/lib/model/confidence_weighted.test.js
@@ -6,18 +6,34 @@ import { ConfidenceWeighted, SoftConfidenceWeighted } from '../../../lib/model/c
 
 import { accuracy } from '../../../lib/evaluate/classification.js'
 
-test('ConfidenceWeighted', () => {
-	const model = new ConfidenceWeighted(0.9)
-	const x = Matrix.concat(Matrix.randn(50, 2, 0, 0.2), Matrix.randn(50, 2, 5, 0.2)).toArray()
-	const t = []
-	for (let i = 0; i < x.length; i++) {
-		t[i] = Math.floor(i / 50) * 2 - 1
-	}
-	model.init(x, t)
-	model.fit()
-	const y = model.predict(x)
-	const acc = accuracy(y, t)
-	expect(acc).toBeGreaterThan(0.95)
+describe('ConfidenceWeighted', () => {
+	test('normal eta', () => {
+		const model = new ConfidenceWeighted(0.9)
+		const x = Matrix.concat(Matrix.randn(50, 2, 0, 0.2), Matrix.randn(50, 2, 5, 0.2)).toArray()
+		const t = []
+		for (let i = 0; i < x.length; i++) {
+			t[i] = Math.floor(i / 50) * 2 - 1
+		}
+		model.init(x, t)
+		model.fit()
+		const y = model.predict(x)
+		const acc = accuracy(y, t)
+		expect(acc).toBeGreaterThan(0.95)
+	})
+
+	test.each([1, 0.5, 0])('eta %p', eta => {
+		const model = new ConfidenceWeighted(eta)
+		const x = Matrix.concat(Matrix.randn(50, 2, 0, 0.2), Matrix.randn(50, 2, 5, 0.2)).toArray()
+		const t = []
+		for (let i = 0; i < x.length; i++) {
+			t[i] = Math.floor(i / 50) * 2 - 1
+		}
+		model.init(x, t)
+		model.fit()
+		const y = model.predict(x)
+		const acc = accuracy(y, t)
+		expect(acc).toBeCloseTo(0.5)
+	})
 })
 
 test.each([1, 2])('SoftConfidenceWeighted %d', version => {
diff --git a/tests/lib/model/crf.test.js b/tests/lib/model/crf.test.js
index 00402038..4ade6c59 100644
--- a/tests/lib/model/crf.test.js
+++ b/tests/lib/model/crf.test.js
@@ -20,3 +20,21 @@ test('fit', () => {
 	expect(prob).toBeGreaterThan(0.4)
 	expect(prob).toBeLessThanOrEqual(1)
 })
+
+test('fit unknown predict label', () => {
+	const model = new CRF()
+	const x = [['a', 'b', 'c']]
+	const y = [[2, 0, 1]]
+
+	for (let i = 0; i < 100; i++) {
+		model.fit(x, y)
+	}
+
+	const tx = [['a', 'd', 'c']]
+	const ty = [[2, 0, 1]]
+	const p = model.predict(tx)
+	expect(p).toEqual(ty)
+	const prob = model.probability(tx[0], ty[0])
+	expect(prob).toBeGreaterThan(0.4)
+	expect(prob).toBeLessThanOrEqual(1)
+})
diff --git a/tests/lib/model/cubic_hermite_spline.test.js b/tests/lib/model/cubic_hermite_spline.test.js
index 28d366d8..62427caa 100644
--- a/tests/lib/model/cubic_hermite_spline.test.js
+++ b/tests/lib/model/cubic_hermite_spline.test.js
@@ -3,23 +3,35 @@ import CubicHermiteSpline from '../../../lib/model/cubic_hermite_spline.js'
 
 import { rmse } from '../../../lib/evaluate/regression.js'
 
-test('interpolation', () => {
-	const model = new CubicHermiteSpline(0, 0)
-	const x = Matrix.random(50, 1, -2, 2).value
-	const t = []
-	for (let i = 0; i < x.length; i++) {
-		t[i] = Math.sin(x[i])
-	}
-	model.fit(x, t)
+describe('interpolation', () => {
+	test('default', () => {
+		const model = new CubicHermiteSpline(0, 0)
+		const x = Matrix.random(50, 1, -2, 2).value
+		const t = []
+		for (let i = 0; i < x.length; i++) {
+			t[i] = Math.sin(x[i])
+		}
+		model.fit(x, t)
 
-	const y = model.predict(x)
-	expect(y).toHaveLength(x.length)
-	for (let i = 0; i < y.length; i++) {
-		expect(y[i]).toBeCloseTo(t[i])
-	}
+		const y = model.predict(x)
+		expect(y).toHaveLength(x.length)
+		for (let i = 0; i < y.length; i++) {
+			expect(y[i]).toBeCloseTo(t[i])
+		}
 
-	const x0 = Matrix.random(100, 1, -2, 2).value
-	const y0 = model.predict(x0)
-	const err = rmse(y0, x0.map(Math.sin))
-	expect(err).toBeLessThan(0.1)
+		const x0 = Matrix.random(100, 1, -2, 2).value
+		const y0 = model.predict(x0)
+		const err = rmse(y0, x0.map(Math.sin))
+		expect(err).toBeLessThan(0.1)
+	})
+
+	test('predict second largest value', () => {
+		const model = new CubicHermiteSpline(0, 0)
+		const x = [0, 2]
+		const t = [0, 1]
+		model.fit(x, t)
+
+		const y = model.predict([1])
+		expect(y[0]).toBeCloseTo(0.5)
+	})
 })
diff --git a/tests/lib/model/cubic_interpolation.test.js b/tests/lib/model/cubic_interpolation.test.js
index 5a37351d..a6278f24 100644
--- a/tests/lib/model/cubic_interpolation.test.js
+++ b/tests/lib/model/cubic_interpolation.test.js
@@ -3,23 +3,35 @@ import CubicInterpolation from '../../../lib/model/cubic_interpolation.js'
 
 import { rmse } from '../../../lib/evaluate/regression.js'
 
-test('interpolation', () => {
-	const model = new CubicInterpolation()
-	const x = Matrix.random(50, 1, -2, 2).value
-	const t = []
-	for (let i = 0; i < x.length; i++) {
-		t[i] = Math.sin(x[i])
-	}
-	model.fit(x, t)
+describe('interpolation', () => {
+	test('default', () => {
+		const model = new CubicInterpolation()
+		const x = Matrix.random(50, 1, -2, 2).value
+		const t = []
+		for (let i = 0; i < x.length; i++) {
+			t[i] = Math.sin(x[i])
+		}
+		model.fit(x, t)
 
-	const y = model.predict(x)
-	expect(y).toHaveLength(x.length)
-	for (let i = 0; i < y.length; i++) {
-		expect(y[i]).toBeCloseTo(t[i])
-	}
+		const y = model.predict(x)
+		expect(y).toHaveLength(x.length)
+		for (let i = 0; i < y.length; i++) {
+			expect(y[i]).toBeCloseTo(t[i])
+		}
 
-	const x0 = Matrix.random(100, 1, -2, 2).value
-	const y0 = model.predict(x0)
-	const err = rmse(y0, x0.map(Math.sin))
-	expect(err).toBeLessThan(0.1)
+		const x0 = Matrix.random(100, 1, -2, 2).value
+		const y0 = model.predict(x0)
+		const err = rmse(y0, x0.map(Math.sin))
+		expect(err).toBeLessThan(0.1)
+	})
+
+	test('predict second largest value', () => {
+		const model = new CubicInterpolation()
+		const x = [0, 2]
+		const t = [0, 1]
+		model.fit(x, t)
+
+		const y = model.predict([1])
+		expect(y[0]).toBeCloseTo(0.5)
+	})
 })
diff --git a/tests/lib/model/dann.test.js b/tests/lib/model/dann.test.js
index 35696b68..a8844682 100644
--- a/tests/lib/model/dann.test.js
+++ b/tests/lib/model/dann.test.js
@@ -33,4 +33,17 @@ describe('classification', () => {
 		const acc = accuracy(y, t)
 		expect(acc).toBeGreaterThan(0.95)
 	})
+
+	test('same number of class choice', () => {
+		const model = new DiscriminantAdaptiveNearestNeighbor(2)
+		const x = [
+			[-1, -1],
+			[1, 1],
+		]
+		const t = ['a', 'b']
+
+		model.fit(x, t)
+		const y = model.predict([[-1, -1]])
+		expect(y).toEqual(['a'])
+	})
 })
diff --git a/tests/lib/model/denclue.test.js b/tests/lib/model/denclue.test.js
index 3a5d0398..41bae894 100644
--- a/tests/lib/model/denclue.test.js
+++ b/tests/lib/model/denclue.test.js
@@ -78,4 +78,20 @@ describe('clustering', () => {
 		const ri = randIndex(y, t)
 		expect(ri).toBeGreaterThan(0.9)
 	})
+
+	test('large h', () => {
+		const model = new DENCLUE(1e6)
+		const n = 50
+		const x = [
+			[0, 0],
+			[1, 1],
+		]
+
+		model.init(x)
+		for (let i = 0; i < 10; i++) {
+			model.fit()
+		}
+		const y = model.predict()
+		expect(y).toEqual([-1, -1])
+	})
 })
diff --git a/tests/lib/model/gaussian_process.test.js b/tests/lib/model/gaussian_process.test.js
index 9e0c7b09..19e45377 100644
--- a/tests/lib/model/gaussian_process.test.js
+++ b/tests/lib/model/gaussian_process.test.js
@@ -6,18 +6,36 @@ import GaussianProcess from '../../../lib/model/gaussian_process.js'
 
 import { rmse } from '../../../lib/evaluate/regression.js'
 
-test('regression', () => {
-	const model = new GaussianProcess()
-	const x = Matrix.random(50, 2, -2, 2).toArray()
-	const t = []
-	for (let i = 0; i < x.length; i++) {
-		t[i] = [x[i][0] + x[i][1] + (Math.random() - 0.5) / 10]
-	}
-	model.init(x, t)
-	for (let i = 0; i < 10; i++) {
-		model.fit(0.001)
-	}
-	const y = model.predict(x)
-	const err = rmse(y, t)[0]
-	expect(err).toBeLessThan(0.5)
+describe('regression', () => {
+	test('fit default', () => {
+		const model = new GaussianProcess(undefined, 10)
+		const x = [
+			[0, 0],
+			[1, 1],
+		]
+		const t = [[0], [1]]
+		model.init(x, t)
+		for (let i = 0; i < 10; i++) {
+			model.fit()
+		}
+		const y = model.predict(x)
+		const err = rmse(y, t)[0]
+		expect(err).toBeLessThan(0.5)
+	})
+
+	test('fit with param', () => {
+		const model = new GaussianProcess()
+		const x = Matrix.random(50, 2, -2, 2).toArray()
+		const t = []
+		for (let i = 0; i < x.length; i++) {
+			t[i] = [x[i][0] + x[i][1] + (Math.random() - 0.5) / 10]
+		}
+		model.init(x, t)
+		for (let i = 0; i < 10; i++) {
+			model.fit(0.001)
+		}
+		const y = model.predict(x)
+		const err = rmse(y, t)[0]
+		expect(err).toBeLessThan(0.5)
+	})
 })
diff --git a/tests/lib/model/gmm.test.js b/tests/lib/model/gmm.test.js
index b4b27f4f..bc190c28 100644
--- a/tests/lib/model/gmm.test.js
+++ b/tests/lib/model/gmm.test.js
@@ -34,6 +34,22 @@ describe('clustering', () => {
 		expect(ri).toBeGreaterThan(0.9)
 	})
 
+	test('clear', () => {
+		const model = new GMM()
+		const x = [
+			[0, 0],
+			[1, 1],
+		]
+		model.add()
+		model.add()
+		for (let i = 0; i < 100; i++) {
+			model.fit(x)
+		}
+		expect(model._k).toBe(2)
+		model.clear()
+		expect(model._k).toBe(0)
+	})
+
 	test('probability', () => {
 		const model = new GMM()
 		const n = 50
@@ -55,22 +71,65 @@ describe('clustering', () => {
 	})
 })
 
-test('regression', () => {
-	const model = new GMR()
-	const x = Matrix.randn(50, 2, 0, 5).toArray()
-	const t = []
-	for (let i = 0; i < x.length; i++) {
-		t[i] = [x[i][0] + x[i][1] + (Math.random() - 0.5) / 10]
-	}
-	model.add()
-	model.add()
-	model.add()
-	for (let i = 0; i < 20; i++) {
+describe('regression', () => {
+	test('predict', () => {
+		const model = new GMR()
+		const x = Matrix.randn(50, 2, 0, 5).toArray()
+		const t = []
+		for (let i = 0; i < x.length; i++) {
+			t[i] = [x[i][0] + x[i][1] + (Math.random() - 0.5) / 10]
+		}
+		model.add()
+		model.add()
+		model.add()
+		for (let i = 0; i < 20; i++) {
+			model.fit(x, t)
+		}
+		const y = model.predict(x)
+		const err = rmse(y, t)[0]
+		expect(err).toBeLessThan(0.5)
+	})
+
+	test('clear', () => {
+		const model = new GMR()
+		const x = [
+			[0, 0],
+			[1, 1],
+		]
+		const t = [[0], [1]]
+		model.add()
+		model.add()
 		model.fit(x, t)
-	}
-	const y = model.predict(x)
-	const err = rmse(y, t)[0]
-	expect(err).toBeLessThan(0.5)
+		expect(model._k).toBe(2)
+		model.clear()
+		expect(model._k).toBe(0)
+	})
+
+	test('probability', () => {
+		const model = new GMR()
+		const x = [
+			[0, 0],
+			[1, 1],
+		]
+		const t = [[0], [1]]
+
+		model.add()
+		model.add()
+		for (let i = 0; i < 100; i++) {
+			model.fit(x, t)
+		}
+
+		const p = model.probability(
+			[
+				[0, 0],
+				[-1, -1],
+			],
+			[[0], [-1]]
+		)
+		for (let c = 0; c < 2; c++) {
+			expect(p[0][c]).toBeGreaterThan(p[1][c])
+		}
+	})
 })
 
 test('semi-classifier', () => {
diff --git a/tests/lib/model/hmm.test.js b/tests/lib/model/hmm.test.js
index 937bafdd..4d2203b1 100644
--- a/tests/lib/model/hmm.test.js
+++ b/tests/lib/model/hmm.test.js
@@ -97,4 +97,17 @@ describe('continuous hmm', () => {
 		expect(mean).toBeCloseTo(0, 1)
 		expect(vari).toBeCloseTo(1, 0)
 	})
+
+	test('generate default', () => {
+		const model = new ContinuousHMM(5)
+		const x = Tensor.randn([10, 7, 2]).toArray()
+		model.fit(x, true)
+
+		const gen = model.generate()
+		expect(gen).toHaveLength(1)
+
+		for (let i = 0; i < gen.length; i++) {
+			expect(gen[i]).toHaveLength(5)
+		}
+	})
 })
diff --git a/tests/lib/model/hopfield.test.js b/tests/lib/model/hopfield.test.js
index 87765366..79d775fd 100644
--- a/tests/lib/model/hopfield.test.js
+++ b/tests/lib/model/hopfield.test.js
@@ -16,3 +16,13 @@ test('reconstruct', () => {
 	const y2 = model.predict([-1, 1, -1, 1, 1, 1])
 	expect(y2).toEqual(x2)
 })
+
+test('reconstruct one', () => {
+	const model = new Hopfield()
+	const x = [1]
+	model.fit([x])
+	const energy = model.energy(x)
+	expect(energy).toBe(0)
+	const y = model.predict(x)
+	expect(y).toEqual([1])
+})
diff --git a/tests/lib/model/ica.test.js b/tests/lib/model/ica.test.js
index 19383d08..82f7bf6d 100644
--- a/tests/lib/model/ica.test.js
+++ b/tests/lib/model/ica.test.js
@@ -34,4 +34,16 @@ describe('dimensionality reduction', () => {
 		const q = coRankingMatrix(x, y, 30, 20)
 		expect(q).toBeGreaterThan(0.9)
 	})
+
+	test('fit twice', () => {
+		const x = [
+			[0, 0],
+			[1, 1],
+		]
+		const model = new ICA()
+		model.fit(x)
+		model.fit(x)
+		const y = model.predict(x)
+		expect(y[0]).toHaveLength(2)
+	})
 })
diff --git a/tests/lib/model/iknn.test.js b/tests/lib/model/iknn.test.js
index 61c520b8..33b12fdb 100644
--- a/tests/lib/model/iknn.test.js
+++ b/tests/lib/model/iknn.test.js
@@ -17,3 +17,19 @@ test('predict', () => {
 	const acc = accuracy(y, t)
 	expect(acc).toBeGreaterThan(0.9)
 })
+
+test('same number of class choice', () => {
+	const model = new IKNN(2, 2)
+	const x = [
+		[-1, -1],
+		[1, 1],
+	]
+	const t = ['a', 'b']
+
+	model.fit(x, t)
+	const y = model.predict([
+		[-1, -1],
+		[1, 1],
+	])
+	expect(y).toEqual(['a', 'b'])
+})
diff --git a/tests/lib/model/incremental_pca.test.js b/tests/lib/model/incremental_pca.test.js
index bc672a67..0b40bcd4 100644
--- a/tests/lib/model/incremental_pca.test.js
+++ b/tests/lib/model/incremental_pca.test.js
@@ -25,4 +25,17 @@ describe('dimensionality reduction', () => {
 		const q = coRankingMatrix(x, y, 30, 20)
 		expect(q).toBeGreaterThan(0.9)
 	})
+
+	test('fit twice', () => {
+		const x = [
+			[0, 0],
+			[1, 1],
+		]
+		const model = new IncrementalPCA()
+		model.fit(x)
+		model.fit(x)
+
+		const y = model.predict(x)
+		expect(y[0]).toHaveLength(2)
+	})
 })
diff --git a/tests/lib/model/isotonic.test.js b/tests/lib/model/isotonic.test.js
index 5cc787a4..e6d6aa8f 100644
--- a/tests/lib/model/isotonic.test.js
+++ b/tests/lib/model/isotonic.test.js
@@ -15,3 +15,13 @@ test('fit', () => {
 	const err = rmse(y, t)
 	expect(err).toBeLessThan(0.5)
 })
+
+test('fit c not match', () => {
+	const model = new IsotonicRegression()
+	const x = [0.0, 1.2, 1.5, 1.1]
+	const t = [0.0, 1.4, 1.3, 1.6]
+	model.fit(x, t)
+	const y = model.predict(x)
+	const err = rmse(y, t)
+	expect(err).toBeLessThan(1.0)
+})
diff --git a/tests/lib/model/kalman_filter.test.js b/tests/lib/model/kalman_filter.test.js
index e9651652..649b5eab 100644
--- a/tests/lib/model/kalman_filter.test.js
+++ b/tests/lib/model/kalman_filter.test.js
@@ -2,16 +2,33 @@ import KalmanFilter from '../../../lib/model/kalman_filter.js'
 
 import { rmse } from '../../../lib/evaluate/regression.js'
 
-test('smoothing', () => {
-	const x = []
-	const t = []
-	for (let i = 0; i < 100; i++) {
-		x[i] = [Math.sin(i / 50) + (Math.random() - 0.5) / 2]
-		t[i] = [Math.sin(i / 50)]
-	}
-	const model = new KalmanFilter()
-	const y = model.fit(x)
-	expect(y).toHaveLength(t.length)
-	const err = rmse(y, t)[0]
-	expect(err).toBeLessThan(rmse(x, t)[0])
+describe('smoothing', () => {
+	test('small dim', () => {
+		const x = []
+		const t = []
+		for (let i = 0; i < 100; i++) {
+			x[i] = [Math.sin(i / 50) + (Math.random() - 0.5) / 2]
+			t[i] = [Math.sin(i / 50)]
+		}
+		const model = new KalmanFilter()
+		const y = model.fit(x)
+		expect(y).toHaveLength(t.length)
+		const err = rmse(y, t)[0]
+		expect(err).toBeLessThan(rmse(x, t)[0])
+	})
+
+	test('large dim', () => {
+		const x = []
+		const t = []
+		for (let i = 0; i < 100; i++) {
+			const v = Math.sin(i / 50) + (Math.random() - 0.5) / 2
+			x[i] = Array.from({ length: 12 }, () => v)
+			t[i] = [Math.sin(i / 50)]
+		}
+		const model = new KalmanFilter()
+		const y = model.fit(x)
+		expect(y).toHaveLength(t.length)
+		const err = rmse(y, t)[0]
+		expect(err).toBeLessThan(Infinity)
+	})
 })
diff --git a/tests/lib/model/knearestneighbor.test.js b/tests/lib/model/knearestneighbor.test.js
index e81d9330..ed4f41ad 100644
--- a/tests/lib/model/knearestneighbor.test.js
+++ b/tests/lib/model/knearestneighbor.test.js
@@ -33,6 +33,22 @@ describe.each([
 		const acc = accuracy(y, t)
 		expect(acc).toBeGreaterThan(0.95)
 	})
+
+	test('same number of class choice', () => {
+		const model = new KNN(2)
+		const x = [
+			[-1, -1],
+			[1, 1],
+		]
+		const t = ['a', 'b']
+
+		model.fit(x, t)
+		const y = model.predict([
+			[-1, -1],
+			[1, 1],
+		])
+		expect(y).toEqual(['a', 'b'])
+	})
 })
 
 describe.each([
diff --git a/tests/lib/model/ksvd.test.js b/tests/lib/model/ksvd.test.js
index 16c03c39..2d260420 100644
--- a/tests/lib/model/ksvd.test.js
+++ b/tests/lib/model/ksvd.test.js
@@ -17,3 +17,12 @@ test('dimensionality reduction', () => {
 	const q = coRankingMatrix(x, y, 30, 20)
 	expect(q).toBeGreaterThan(0.9)
 })
+
+test('dimensionality reduction small norm', () => {
+	const x = [[0, 0, 0, 0, 0]]
+	const model = new KSVD(x, 2)
+
+	model.fit()
+	const y = model.predict()
+	expect(y).toEqual([[0, 0]])
+})
diff --git a/tests/lib/model/label_propagation.test.js b/tests/lib/model/label_propagation.test.js
index 9025b8c7..80be4ce8 100644
--- a/tests/lib/model/label_propagation.test.js
+++ b/tests/lib/model/label_propagation.test.js
@@ -22,3 +22,23 @@ test.each([undefined, 'rbf', { name: 'rbf', sigma: 0.2 }, { name: 'knn', k: 10 }
 	const acc = accuracy(y, t_org)
 	expect(acc).toBeGreaterThan(0.95)
 })
+
+test('semi-classifier k=0', () => {
+	const model = new LabelPropagation({ name: 'rbf', sigma: 0.1, k: 0 })
+	const x = Matrix.concat(Matrix.randn(50, 2, 0, 0.2), Matrix.randn(50, 2, 5, 0.2)).toArray()
+	const t = []
+	const t_org = []
+	for (let i = 0; i < x.length; i++) {
+		t_org[i] = t[i] = String.fromCharCode('a'.charCodeAt(0) + Math.floor(i / 50))
+		if (Math.random() < 0.5) {
+			t[i] = null
+		}
+	}
+	model.init(x, t)
+	for (let i = 0; i < 20; i++) {
+		model.fit()
+	}
+	const y = model.predict(x)
+	const acc = accuracy(y, t_org)
+	expect(acc).toBeGreaterThan(0.6)
+})
diff --git a/tests/lib/model/label_spreading.test.js b/tests/lib/model/label_spreading.test.js
index e21c51b2..1339a254 100644
--- a/tests/lib/model/label_spreading.test.js
+++ b/tests/lib/model/label_spreading.test.js
@@ -26,3 +26,23 @@ test.each([
 	const acc = accuracy(y, t_org)
 	expect(acc).toBeGreaterThan(0.95)
 })
+
+test('semi-classifier k=0', () => {
+	const model = new LabelSpreading(0.5, { name: 'rbf', sigma: 0.1, k: 0 })
+	const x = Matrix.concat(Matrix.randn(50, 2, 0, 0.2), Matrix.randn(50, 2, 5, 0.2)).toArray()
+	const t = []
+	const t_org = []
+	for (let i = 0; i < x.length; i++) {
+		t_org[i] = t[i] = String.fromCharCode('a'.charCodeAt(0) + Math.floor(i / 50))
+		if (Math.random() < 0.5) {
+			t[i] = null
+		}
+	}
+	model.init(x, t)
+	for (let i = 0; i < 20; i++) {
+		model.fit()
+	}
+	const y = model.predict(x)
+	const acc = accuracy(y, t_org)
+	expect(acc).toBeGreaterThanOrEqual(0.5)
+})
diff --git a/tests/lib/model/ladder_network.test.js b/tests/lib/model/ladder_network.test.js
index 74f6ff90..95332592 100644
--- a/tests/lib/model/ladder_network.test.js
+++ b/tests/lib/model/ladder_network.test.js
@@ -26,3 +26,17 @@ test('semi-classifier', () => {
 	const acc = accuracy(y, t_org)
 	expect(acc).toBeGreaterThan(0.95)
 })
+
+test('semi-classifier learn only', () => {
+	const model = new LadderNetwork([5], [0.001, 0.0001, 0.0001], 'tanh', 'adam')
+	const x = [
+		[0, 0],
+		[1, 1],
+	]
+	const t = ['a', 'a']
+	model.fit([[0, 0]], ['a'], 1, 0.001)
+	model.fit([[1, 1]], [null], 1, 0.001)
+	const y = model.predict(x)
+	const acc = accuracy(y, t)
+	expect(acc).toBeGreaterThan(0.95)
+})
diff --git a/tests/lib/model/laplacian_eigenmaps.test.js b/tests/lib/model/laplacian_eigenmaps.test.js
index 0e75ae10..88922b80 100644
--- a/tests/lib/model/laplacian_eigenmaps.test.js
+++ b/tests/lib/model/laplacian_eigenmaps.test.js
@@ -17,3 +17,13 @@ describe.each([undefined, 'knn', { name: 'rbf' }])('dimensionality reduction aff
 		expect(q).toBeGreaterThan(0.9)
 	})
 })
+
+test('dimensionality k=0', () => {
+	const x = Matrix.concat(Matrix.randn(30, 5, 0, 0.2), Matrix.randn(30, 5, 5, 0.2)).toArray()
+	const model = new LaplacianEigenmaps(2, { name: 'rbf', k: 0 })
+
+	const y = model.predict(x)
+	expect(y[0]).toHaveLength(2)
+	const q = coRankingMatrix(x, y, 20, 20)
+	expect(q).toBeGreaterThan(0.7)
+})
diff --git a/tests/lib/model/lmclus.test.js b/tests/lib/model/lmclus.test.js
index 626b6251..eb532192 100644
--- a/tests/lib/model/lmclus.test.js
+++ b/tests/lib/model/lmclus.test.js
@@ -15,6 +15,7 @@ test('clustering', () => {
 	).toArray()
 
 	model.fit(x)
+	expect(model.size).toBeGreaterThanOrEqual(3)
 	const y = model.predict()
 	expect(y).toHaveLength(x.length)
 
diff --git a/tests/lib/model/markov_switching.test.js b/tests/lib/model/markov_switching.test.js
index aff40d76..f68355db 100644
--- a/tests/lib/model/markov_switching.test.js
+++ b/tests/lib/model/markov_switching.test.js
@@ -33,3 +33,14 @@ test('anomaly detection', () => {
 	expect(Math.abs(prob[0][0] - prob[0][1])).toBeCloseTo(1)
 	expect(prob[0][0] - prob[0][1]).toBeCloseTo(prob[1][1] - prob[1][0])
 })
+
+test('anomaly detection fit twice', () => {
+	const model = new MarkovSwitching(2)
+	const n = 50
+	const x = Matrix.concat(Matrix.random(n, 2, 0, 1), Matrix.random(n, 2, 2, 3)).toArray()
+
+	model.fit(x, 1, 1)
+	model.fit(x, 1, 1)
+	const p = model.predict(x)
+	expect(p).toHaveLength(99)
+})
diff --git a/tests/lib/model/mars.test.js b/tests/lib/model/mars.test.js
index 0718c61a..9c03ef1a 100644
--- a/tests/lib/model/mars.test.js
+++ b/tests/lib/model/mars.test.js
@@ -15,3 +15,33 @@ test('fit', () => {
 	const err = rmse(y, t)[0]
 	expect(err).toBeLessThan(0.5)
 })
+
+test('fit update best lof', () => {
+	const model = new MARS(20)
+	const x = [
+		[0.4946112861933446, -0.24979407544789328],
+		[0.6710977380517686, 2.760557507387143],
+		[-0.5451316855831966, -1.7861984864750804],
+		[1.5413600325044114, 0.6113068721545195],
+		[-0.15915692646845286, 0.36163408767347377],
+		[-4.456221847339759, 2.6448274560926066],
+		[0.9859756767948494, -0.7623872449215827],
+		[1.6964768061759719, 4.561617731454047],
+		[-3.0039970134614746, 3.7582323314173856],
+	]
+	const t = [
+		[5.0135998349450785],
+		[8.344300706946722],
+		[2.8613517704846276],
+		[6.965641204443869],
+		[5.4431303914749085],
+		[3.072585145244251],
+		[5.404783275862362],
+		[11.42125315864682],
+		[5.875750792935508],
+	]
+	model.fit(x, t)
+	const y = model.predict(x)
+	const err = rmse(y, t)[0]
+	expect(err).toBeLessThan(0.5)
+})
diff --git a/tests/lib/model/mlle.test.js b/tests/lib/model/mlle.test.js
index 34cbcbb3..79bdab14 100644
--- a/tests/lib/model/mlle.test.js
+++ b/tests/lib/model/mlle.test.js
@@ -22,4 +22,22 @@ describe('dimensionality reduction', () => {
 		const q = coRankingMatrix(x, y, 20, 30)
 		expect(q).toBeGreaterThan(0.9)
 	})
+
+	test('odd size and less than eta', () => {
+		const x = [
+			[-0.48, -0.18, -0.14],
+			[0.44, -0.35, 0.08],
+			[-0.15, 0.76, 0.11],
+			[-0.52, -0.33, 0.12],
+			[5.51, 5.27, 5.07],
+			[4.63, 4.63, 5.03],
+			[4.55, 4.58, 4.66],
+			[4.7, 4.5, 5.5],
+			[5.53, 4.53, 4.77],
+		]
+		const y = new MLLE(8).predict(x)
+		expect(y[0]).toHaveLength(3)
+		const q = coRankingMatrix(x, y, 4, 5)
+		expect(q).toBeGreaterThan(0.95)
+	})
 })
diff --git a/tests/lib/model/mod.test.js b/tests/lib/model/mod.test.js
index 2b36c251..95c8be9a 100644
--- a/tests/lib/model/mod.test.js
+++ b/tests/lib/model/mod.test.js
@@ -17,3 +17,12 @@ test('dimensionality reduction', () => {
 	const q = coRankingMatrix(x, y, 30, 20)
 	expect(q).toBeGreaterThan(0.9)
 })
+
+test('dimensionality reduction small norm', () => {
+	const x = [[0, 0, 0, 0, 0]]
+	const model = new MOD(x, 2)
+
+	model.fit()
+	const y = model.predict()
+	expect(y).toEqual([[0, 0]])
+})
diff --git a/tests/lib/model/optics.test.js b/tests/lib/model/optics.test.js
index 33c1d798..adec0de6 100644
--- a/tests/lib/model/optics.test.js
+++ b/tests/lib/model/optics.test.js
@@ -46,4 +46,14 @@ describe('clustering', () => {
 		const ri = randIndex(y, t)
 		expect(ri).toBeGreaterThan(0.9)
 	})
+
+	test('large minpts', () => {
+		const model = new OPTICS(0.4, 0, 100)
+		const n = 100
+		const x = Matrix.concat(Matrix.randn(n, 2, 0, 0.1), Matrix.randn(n, 2, 5, 0.1)).toArray()
+
+		model.fit(x)
+		const y = model.predict()
+		expect(y).toHaveLength(x.length)
+	})
 })
diff --git a/tests/lib/model/pegasos.test.js b/tests/lib/model/pegasos.test.js
index e6e478de..f1363443 100644
--- a/tests/lib/model/pegasos.test.js
+++ b/tests/lib/model/pegasos.test.js
@@ -18,3 +18,20 @@ test('fit', () => {
 	const acc = accuracy(y, t)
 	expect(acc).toBeGreaterThan(0.9)
 })
+
+test('fit do projection', () => {
+	const model = new Pegasos(0.05)
+	model._do_projection = true
+	const x = Matrix.concat(Matrix.randn(50, 2, 0, 0.2), Matrix.randn(50, 2, 5, 0.2)).toArray()
+	const t = []
+	for (let i = 0; i < x.length; i++) {
+		t[i] = Math.floor(i / 50) * 2 - 1
+	}
+	model.init(x, t)
+	for (let i = 0; i < 100; i++) {
+		model.fit()
+	}
+	const y = model.predict(x)
+	const acc = accuracy(y, t)
+	expect(acc).toBeGreaterThan(0.9)
+})
diff --git a/tests/lib/model/pls.test.js b/tests/lib/model/pls.test.js
index f6408f8e..3febd6bf 100644
--- a/tests/lib/model/pls.test.js
+++ b/tests/lib/model/pls.test.js
@@ -12,7 +12,7 @@ describe('regression', () => {
 			t[i] = [x[i][0] + x[i][1] + (Math.random() - 0.5) / 100]
 		}
 		model.init(x, t)
-		model.fit(x, t)
+		model.fit()
 
 		const y = model.predict(x)
 		const err = rmse(y, t)[0]
@@ -27,6 +27,6 @@ describe('regression', () => {
 			t[i] = [0, 0]
 		}
 		model.init(x, t)
-		expect(() => model.fit(x, t)).toThrow()
+		expect(() => model.fit()).toThrow()
 	})
 })
diff --git a/tests/lib/model/rbm.test.js b/tests/lib/model/rbm.test.js
index a9552632..2c8aabdc 100644
--- a/tests/lib/model/rbm.test.js
+++ b/tests/lib/model/rbm.test.js
@@ -4,24 +4,47 @@ jest.retryTimes(5)
 import Matrix from '../../../lib/util/matrix.js'
 import { RBM, GBRBM } from '../../../lib/model/rbm.js'
 
-test('reconstruct RBM', () => {
-	const model = new RBM(10)
-	const x = [
-		[1, 1, 1, 1, 0, 0, 0],
-		[0, 0, 0, 0, 1, 1, 1],
-	]
-	model.fit(x)
-	const oldenergy = model.energy(x[0], model._h([x[0]], false)[0])
-	for (let i = 0; i < 10000; i++) {
+describe('reconstruct RBM', () => {
+	test('default', () => {
+		const model = new RBM(10)
+		const x = [
+			[1, 1, 1, 1, 0, 0, 0],
+			[0, 0, 0, 0, 1, 1, 1],
+		]
 		model.fit(x)
-	}
-	const newenergy = model.energy(x[0], model._h([x[0]], false)[0])
-	expect(newenergy).toBeLessThanOrEqual(oldenergy)
-	const y = model.predict([
-		[1, 0, 1, 1, 0, 0, 0],
-		[0, 1, 0, 0, 1, 1, 1],
-	])
-	expect(y).toEqual(x)
+		const oldenergy = model.energy(x[0], model._h([x[0]], false)[0])
+		for (let i = 0; i < 10000; i++) {
+			model.fit(x)
+		}
+		const newenergy = model.energy(x[0], model._h([x[0]], false)[0])
+		expect(newenergy).toBeLessThanOrEqual(oldenergy)
+		const y = model.predict([
+			[1, 0, 1, 1, 0, 0, 0],
+			[0, 1, 0, 0, 1, 1, 1],
+		])
+		expect(y).toEqual(x)
+	})
+
+	test('k=2', () => {
+		const model = new RBM(10)
+		model._k = 2
+		const x = [
+			[1, 1, 1, 1, 0, 0, 0],
+			[0, 0, 0, 0, 1, 1, 1],
+		]
+		model.fit(x)
+		const oldenergy = model.energy(x[0], model._h([x[0]], false)[0])
+		for (let i = 0; i < 10000; i++) {
+			model.fit(x)
+		}
+		const newenergy = model.energy(x[0], model._h([x[0]], false)[0])
+		expect(newenergy).toBeLessThanOrEqual(oldenergy)
+		const y = model.predict([
+			[1, 0, 1, 1, 0, 0, 0],
+			[0, 1, 0, 0, 1, 1, 1],
+		])
+		expect(y).toEqual(x)
+	})
 })
 
 test('reconstruct GBRBM', () => {
diff --git a/tests/lib/model/sdar.test.js b/tests/lib/model/sdar.test.js
index d2e4539d..54abad57 100644
--- a/tests/lib/model/sdar.test.js
+++ b/tests/lib/model/sdar.test.js
@@ -1,23 +1,38 @@
 import SDAR from '../../../lib/model/sdar.js'
 
-test('linear', () => {
-	const model = new SDAR(10, 0.1)
-	const x = []
-	for (let i = 0; i < 1000; i++) {
-		x[i] = i / 10
-	}
+describe('probability', () => {
+	test('linear', () => {
+		const model = new SDAR(10, 0.1)
+		const x = []
+		for (let i = 0; i < 100; i++) {
+			x[i] = i / 10
+		}
 
-	const future = model.predict(x, 20)
-	expect(future).toHaveLength(20)
+		const prob = model.probability(x)
+		expect(prob).toHaveLength(100)
+	})
 })
 
-test('sin', () => {
-	const model = new SDAR(50)
-	const x = []
-	for (let i = 0; i < 1000; i++) {
-		x[i] = Math.sin(i / 10)
-	}
+describe('predict', () => {
+	test('linear', () => {
+		const model = new SDAR(10, 0.1)
+		const x = []
+		for (let i = 0; i < 1000; i++) {
+			x[i] = i / 10
+		}
 
-	const future = model.predict(x, 50)
-	expect(future).toHaveLength(50)
+		const future = model.predict(x, 20)
+		expect(future).toHaveLength(20)
+	})
+
+	test('sin', () => {
+		const model = new SDAR(50)
+		const x = []
+		for (let i = 0; i < 1000; i++) {
+			x[i] = Math.sin(i / 10)
+		}
+
+		const future = model.predict(x, 50)
+		expect(future).toHaveLength(50)
+	})
 })
diff --git a/tests/lib/model/self_training.test.js b/tests/lib/model/self_training.test.js
index 5aacb571..48ed5cf4 100644
--- a/tests/lib/model/self_training.test.js
+++ b/tests/lib/model/self_training.test.js
@@ -46,3 +46,27 @@ test('semi-classifier', () => {
 	const acc = accuracy(y, t_org)
 	expect(acc).toBeGreaterThan(0.95)
 })
+
+test('semi-classifier', () => {
+	const dt = new DecisionTreeClassifier('CART')
+	const model = new SelfTraining(
+		{
+			fit(x, y) {},
+			predict(x) {
+				return x.map(() => ({ category: 0, score: 0 }))
+			},
+		},
+		0.5
+	)
+	const x = [
+		[0, 0],
+		[1, 1],
+	]
+	const t = [0, null]
+	model.init(x, t)
+	for (let i = 0; i < 1; i++) {
+		model.fit()
+	}
+	const y = model.predict(x)
+	expect(y).toEqual([0, null])
+})
diff --git a/tests/lib/model/semi_supervised_naive_bayes.test.js b/tests/lib/model/semi_supervised_naive_bayes.test.js
index 2185b8ca..b352d87c 100644
--- a/tests/lib/model/semi_supervised_naive_bayes.test.js
+++ b/tests/lib/model/semi_supervised_naive_bayes.test.js
@@ -36,3 +36,15 @@ test('semi-classifier', () => {
 	const acc = accuracy(y, t_org)
 	expect(acc).toBeGreaterThan(0.95)
 })
+
+test('semi-classifier unknown data', () => {
+	const model = new SemiSupervisedNaiveBayes()
+	const x = [['May', 'I', 'have', 'a', 'large', 'container', 'of', 'coffee']]
+	const t = ['a']
+
+	model.init(x, t)
+	model.fit()
+
+	const y = model.predict([['Dummy']])
+	expect(y).toEqual([null])
+})
diff --git a/tests/lib/model/smirnov_grubbs.test.js b/tests/lib/model/smirnov_grubbs.test.js
index 2d8a9f55..1f031ed6 100644
--- a/tests/lib/model/smirnov_grubbs.test.js
+++ b/tests/lib/model/smirnov_grubbs.test.js
@@ -19,4 +19,52 @@ describe('anomaly detection', () => {
 		const y = model.predict(x)
 		expect(y).toEqual([false, false])
 	})
+
+	test('no outlier', () => {
+		const model = new SmirnovGrubbs(1)
+		const x = []
+		for (let i = 0; i < 100; i++) {
+			for (let j = 0; j < 100; j++) {
+				x.push([i / 100, j / 100])
+			}
+		}
+		const y = model.predict(x)
+		for (let i = 0; i < y.length; i++) {
+			expect(y[i]).toBe(false)
+		}
+	})
+
+	test('p is 1', () => {
+		const model = new SmirnovGrubbs(3)
+		const x = [
+			[0, 0],
+			[1, 1],
+			[2, 2],
+		]
+		const y = model.predict(x)
+		expect(y).toEqual([false, false, false])
+	})
+
+	test('p is over 0.5', () => {
+		const model = new SmirnovGrubbs(2)
+		const x = [
+			[0, 0],
+			[1, 1],
+			[2, 2],
+		]
+		const y = model.predict(x)
+		expect(y).toEqual([true, false, false])
+	})
+
+	test('n is 2', () => {
+		const model = new SmirnovGrubbs(2)
+		const x = [
+			[0, 0],
+			[1, 1],
+			[2, 2],
+			[3, 3],
+		]
+		const y = model.predict(x)
+		expect(y).toEqual([true, false, false, false])
+	})
 })
diff --git a/tests/lib/model/snakes.test.js b/tests/lib/model/snakes.test.js
index 688a8d3d..7d4e1d4f 100644
--- a/tests/lib/model/snakes.test.js
+++ b/tests/lib/model/snakes.test.js
@@ -1,38 +1,72 @@
 import Tensor from '../../../lib/util/tensor.js'
 import Snakes from '../../../lib/model/snakes.js'
 
-test('predict', () => {
-	const model = new Snakes(1, 1, 1, 12)
-	const n = 100
-	const x = Tensor.zeros([n, n, 1]).toArray()
-	for (let i = 0; i < n; i++) {
-		for (let j = 0; j < n; j++) {
-			if ((i - n / 2) ** 2 + (j - n / 2) ** 2 < (n / 4) ** 2) {
-				x[i][j][0] = 255
+describe('predict', () => {
+	test('default', () => {
+		const model = new Snakes(1, 1, 1)
+		const n = 10
+		const x = Tensor.zeros([n, n, 1]).toArray()
+		for (let i = 0; i < n; i++) {
+			for (let j = 0; j < n; j++) {
+				if ((i - n / 2) ** 2 + (j - n / 2) ** 2 < (n / 4) ** 2) {
+					x[i][j][0] = 255
+				}
 			}
 		}
-	}
 
-	model.init(x)
-	for (let i = 0; i < 100; i++) {
+		model.init(x)
 		model.fit()
-	}
 
-	const y = model.predict(x)
-	const b = [0, 0]
-	for (let i = 0; i < n; i++) {
-		for (let j = 0; j < n; j++) {
-			const r2 = (i - n / 2) ** 2 + (j - n / 2) ** 2
-			if ((n / 4 - 2) ** 2 < r2 && r2 < (n / 4 + 2) ** 2) {
-				if (y[i][j]) {
-					b[0]++
-				} else {
-					b[1]++
+		const y = model.predict(x)
+		const b = [0, 0]
+		for (let i = 0; i < n; i++) {
+			for (let j = 0; j < n; j++) {
+				const r2 = (i - n / 2) ** 2 + (j - n / 2) ** 2
+				if ((n / 4 - 2) ** 2 < r2 && r2 < (n / 4 + 2) ** 2) {
+					if (y[i][j]) {
+						b[0]++
+					} else {
+						b[1]++
+					}
+				}
+			}
+		}
+		expect(b[0] / (b[0] + b[1])).toBeGreaterThan(0.1)
+	})
+
+	test('with params', () => {
+		const model = new Snakes(1, 1, 1, 12)
+		const n = 100
+		const x = Tensor.zeros([n, n, 1]).toArray()
+		for (let i = 0; i < n; i++) {
+			for (let j = 0; j < n; j++) {
+				if ((i - n / 2) ** 2 + (j - n / 2) ** 2 < (n / 4) ** 2) {
+					x[i][j][0] = 255
+				}
+			}
+		}
+
+		model.init(x)
+		for (let i = 0; i < 100; i++) {
+			model.fit()
+		}
+
+		const y = model.predict(x)
+		const b = [0, 0]
+		for (let i = 0; i < n; i++) {
+			for (let j = 0; j < n; j++) {
+				const r2 = (i - n / 2) ** 2 + (j - n / 2) ** 2
+				if ((n / 4 - 2) ** 2 < r2 && r2 < (n / 4 + 2) ** 2) {
+					if (y[i][j]) {
+						b[0]++
+					} else {
+						b[1]++
+					}
+					continue
 				}
-				continue
+				expect(y[i][j]).toBe(false)
 			}
-			expect(y[i][j]).toBe(false)
 		}
-	}
-	expect(b[0] / (b[0] + b[1])).toBeGreaterThan(0.2)
+		expect(b[0] / (b[0] + b[1])).toBeGreaterThan(0.2)
+	})
 })
diff --git a/tests/lib/model/svm.test.js b/tests/lib/model/svm.test.js
index 7ec3bdc6..2cacdbd0 100644
--- a/tests/lib/model/svm.test.js
+++ b/tests/lib/model/svm.test.js
@@ -38,4 +38,20 @@ describe('classification', () => {
 		const acc = accuracy(y.map(Math.sign), t.map(Math.sign))
 		expect(acc).toBeGreaterThan(0.9)
 	})
+
+	test('uniform kernel', () => {
+		const model = new SVM((a, b) => (a.reduce((s, v, i) => s + (v - b[i]) ** 2, 0) < 1 ? 1 : 0))
+		const x = Matrix.concat(Matrix.randn(50, 2, 0, 0.2), Matrix.randn(50, 2, 5, 0.2)).toArray()
+		const t = []
+		for (let i = 0; i < x.length; i++) {
+			t[i] = Math.floor(i / 50) * 2 - 1
+		}
+		model.init(x, t)
+		for (let i = 0; i < 100; i++) {
+			model.fit()
+		}
+		const y = model.predict(x)
+		const acc = accuracy(y.map(Math.sign), t.map(Math.sign))
+		expect(acc).toBeGreaterThan(0.9)
+	})
 })
diff --git a/tests/lib/model/thompson.test.js b/tests/lib/model/thompson.test.js
index db4e1bd2..4a7acdcf 100644
--- a/tests/lib/model/thompson.test.js
+++ b/tests/lib/model/thompson.test.js
@@ -19,4 +19,52 @@ describe('anomaly detection', () => {
 		const y = model.predict(x)
 		expect(y).toEqual([false, false])
 	})
+
+	test('no outlier', () => {
+		const model = new Thompson(1)
+		const x = []
+		for (let i = 0; i < 100; i++) {
+			for (let j = 0; j < 100; j++) {
+				x.push([i / 100, j / 100])
+			}
+		}
+		const y = model.predict(x)
+		for (let i = 0; i < y.length; i++) {
+			expect(y[i]).toBe(false)
+		}
+	})
+
+	test('p is 1', () => {
+		const model = new Thompson(3)
+		const x = [
+			[0, 0],
+			[1, 1],
+			[2, 2],
+		]
+		const y = model.predict(x)
+		expect(y).toEqual([false, false, false])
+	})
+
+	test('p is over 0.5', () => {
+		const model = new Thompson(2)
+		const x = [
+			[0, 0],
+			[1, 1],
+			[2, 2],
+		]
+		const y = model.predict(x)
+		expect(y).toEqual([true, false, false])
+	})
+
+	test('n is 2', () => {
+		const model = new Thompson(2)
+		const x = [
+			[0, 0],
+			[1, 1],
+			[2, 2],
+			[3, 3],
+		]
+		const y = model.predict(x)
+		expect(y).toEqual([true, false, false, false])
+	})
 })
diff --git a/tests/lib/model/tietjen_moore.test.js b/tests/lib/model/tietjen_moore.test.js
index 4099274d..47ab8534 100644
--- a/tests/lib/model/tietjen_moore.test.js
+++ b/tests/lib/model/tietjen_moore.test.js
@@ -19,3 +19,17 @@ test.each([1, 2, 3])('anomaly detection %d', k => {
 	expect(c).toBe(k - 1)
 	expect(y[y.length - 1]).toBe(true)
 })
+
+test('no outlier', () => {
+	const model = new TietjenMoore(1, 0.1)
+	const x = []
+	for (let i = 0; i < 100; i++) {
+		for (let j = 0; j < 100; j++) {
+			x.push([i / 100, j / 100])
+		}
+	}
+	const y = model.predict(x)
+	for (let i = 0; i < y.length; i++) {
+		expect(y[i]).toBe(false)
+	}
+})
diff --git a/tests/lib/model/tsne.test.js b/tests/lib/model/tsne.test.js
index 99663cc4..0ac12b35 100644
--- a/tests/lib/model/tsne.test.js
+++ b/tests/lib/model/tsne.test.js
@@ -19,6 +19,18 @@ describe('SNE dimension reduction', () => {
 		const q = coRankingMatrix(x, y, 10, 10)
 		expect(q).toBeGreaterThan(0.9)
 	})
+
+	test('infinite sigma', () => {
+		const x = [
+			[0, 0],
+			[1, 1],
+		]
+		const model = new SNE(x, 1)
+
+		model.fit()
+		const y = model.predict()
+		expect(y).toHaveLength(2)
+	})
 })
 
 describe('tSNE dimension reduction', () => {
@@ -33,4 +45,16 @@ describe('tSNE dimension reduction', () => {
 		const q = coRankingMatrix(x, y, 10, 10)
 		expect(q).toBeGreaterThan(0.9)
 	})
+
+	test('infinite sigma', () => {
+		const x = [
+			[0, 0],
+			[1, 1],
+		]
+		const model = new tSNE(x, 1)
+
+		model.fit()
+		const y = model.predict()
+		expect(y).toHaveLength(2)
+	})
 })
diff --git a/tests/lib/model/word_to_vec.test.js b/tests/lib/model/word_to_vec.test.js
index e7152b0a..1cdf6523 100644
--- a/tests/lib/model/word_to_vec.test.js
+++ b/tests/lib/model/word_to_vec.test.js
@@ -35,6 +35,19 @@ describe.each(['CBOW', 'skip-gram'])('embedding %s', method => {
 			expect(y[i]).toHaveLength(2)
 		}
 	})
+
+	test('small number of words', () => {
+		const x = ['May', 'I', 'have', 'a', 'large', 'container', 'of', 'coffee']
+		const model = new Word2Vec(method, 1, 3, 2, 'adam')
+
+		model.fit(x, 1, 0.1, 10)
+
+		const y = model.reduce(x)
+		expect(y).toHaveLength(x.length)
+		for (let i = 0; i < y.length; i++) {
+			expect(y[i]).toHaveLength(2)
+		}
+	})
 })
 
 test('predict unknown', () => {